World Nuclear Industry Status Report 2025 (HTML version)

Foreword by

Letizia Magaldi

President, Kyoto Club,
Rome, Italy

By

Mycle Schneider

Independent Analyst,
Paris, France

Project Coordinator and Lead Author

Julie Hazemann

Director of EnerWebWatch,
Paris, France

Documentary Research, Modelling, and Datavisualization

Phred Dvorak

Independent Journalist
Washington D.C., United States

Contributing Author

Emilio Godoy

Independent Investigative Journalist
Mexico

Contributing Author

Dmitry Gorchakov

Nuclear Adviser, Bellona Foundation
Vilnius, Lithuania

Contributing Author

Özgür Gürbüz

Independent Consultant and Researcher
Istanbul, Türkiye

Contributing Author

Bernd Hrdy

Researcher, Institute for Safety and Risk Sciences, University of Natural Resources and Life Sciences (BOKU)
Vienna, Austria

Contributing Author

Paul Jobin

Associate Research Fellow,
Institute of Sociology, Academia Sinica
Taipei, Taiwan

Contributing Author

Timothy Judson

Independent Consultant Syracuse,
New York, United States

Contributing Author

Yuki Kobayashi

Senior Research Fellow, Security Studies Program, Sasakwa Peace Foundation
Tokyo, Japan

Contributing Author

Nikolaus Müllner

Head, Institute for Safety and Risk Sciences, University of Natural Resources and Life Sciences (BOKU)
Chairman, International Nuclear Risk Assessment Group (INRAG)
Vienna, Austria

Contributing Author

M.V. Ramana

Simons Chair in Disarmament, Global and Human Security with the School of Public Policy and Global Affairs (SPPGA), University of British Columbia
Vancouver, Canada

Contributing Author

Ruggero Schleicher-Tappeser

Independent Consultant and Writer in Energy Policies
Berlin, Germany

Contributing Author

Sebastian Stier

European Patent Attorney
Munich, Germany

Contributing Author

Tatsujiro Suzuki

President, Peace Depot Visiting Professor, Research Center for Nuclear Weapons Abolition (RECNA),
Nagasaki University,
Former Vice-Chairman of the Japan Atomic Energy Commission, Japan

Contributing Author

Yun-Chung Ting

Postdoctoral Research Associate, Institute of Sociology, Academia Sinica
Taipei, Taiwan

Contributing Author

Alexander Wimmers

Research Associate at the Workgroup for Economic and Infrastructure Policy (WIP), Berlin University of Technology (TU),
Berlin, Germany

Contributing Author

Hartmut Winkler

Professor, University of Johannesburg,
South Africa

Contributing Author

Maahin Ahmed

Freelance Copyeditor
Calgary, Canada

English Language Copyeditor

Nina Schneider

Proofreader and Translator
Paris, France

Fact-checker, Proofreader, Producer

Agnès Stienne

Artist, Graphic Designer, Cartographer
Le Mans, France

Graphic Design and Layout

Friedhelm Meinass

Visual Artist, Painter
Rodgau, Germany

Cover-page Design and Layout

Foreword

by Letizia Magaldi1

The history of nuclear power in Italy is characterized by a pioneering role in the construction of power plants, which led the country to become the third largest producer of nuclear power in the world in the 1960s, with plants in Latina, Trino, Garigliano, and Caorso. Then, within a few decades, two referendums—in 1987 in the wake of Chornobyl and in 2011 after Fukushima—drew a clear line, sanctioning a voluntary withdrawal from nuclear power, an unprecedented choice in Europe. However, while nuclear power disappeared from the political agenda, the expertise did not vanish, and several companies continued to work abroad, also collaborating on the international ITER2 program.

Today, driven by an increasingly alarming climate crisis, by the European taxonomy that has recognized nuclear power as a sustainable activity provided that strict standards are met, and by the acceleration represented by the energy shock triggered by the war in Ukraine, Italy has returned to questioning the use of nuclear power. Price volatility and the fragility of a system still dependent on gas and electricity imports have highlighted how precarious our energy mix is. The electricity grid, which is unbalanced between the north and south, requires programmable sources capable of compensating for the fluctuating generation of renewables. It is in this context that in 2025 the government opened the door to a possible revival of the sector, aiming to achieve a share of between 11 percent and 22 percent of electricity production by 2050 with nuclear power, in particular with Small Modular Reactor (SMR) and Advanced Modular Reactor (AMR) plants.

The industrial world is moving forward, and the Meloni government is pushing in this direction. Studies such as those by Confindustria–ENEA and EY estimate a market worth €46–50 billion (US$54–58 billion) by 2050, 117,000 new jobs and up to €10 billion (US$11.6 billion) saved each year in energy bills.

In reality, as the history of recent decades in the West shows, these scenarios represent an aspiration that has little chance of being realized.

The Bank of Italy itself, in its Occasional Paper ‘L’atomo fuggente’ (The Fleeing Atom) published in June 2025, highlights how the return of nuclear power is unlikely to have a significant impact on the final price of electricity, since even modular reactors remain highly capital-intensive investments, characterized by returns spread over the long term. The main contribution is therefore not to be found in a reduction in immediate costs, but rather in the ability to stabilize the volatility of electricity prices, mitigating exposure to international markets and offering greater protection to businesses and households in the face of possible shocks. However, the study also highlights new potential critical issues related to dependence on foreign supplies of fuel and advanced components, as well as the geopolitical risks arising from global uranium chains. A further concern relates to timing: the construction times, which are inevitably long even for SMRs, appear difficult to reconcile with the urgency of reducing emissions. From this perspective, if nuclear power ever returns to the national energy mix, it will at best complement renewables in a decarbonization process that will continue to rely heavily on solar, wind, and storage systems.

The risk, if anything, is that renewables’ growth will slow down, reaching levels above 80 percent by the middle of the next decade (Germany’s target for 2030).

An analysis of characteristics of the emerging overall energy system indicates that photovoltaics, batteries, and power electronics are already changing the logic of the overall system in a way that makes it increasingly difficult to integrate nuclear energy. Time frames play a key role: until a nuclear power plant, decided on today, begins to produce electricity, new technologies will have radically transformed the energy system.

Interesting insights into these critical issues are contained in the chapter “Challenges of integrating nuclear energy into the energy system” of the World Nuclear Industry Status Report 2025. It highlights how the logic of the energy system is changing radically, with increasingly decentralized electricity production involving millions of investors and operators. Generation technologies are transforming from gigantic and customized to modular and mass-produced. Moreover, grid-forming inverters are now available that can play an important role in stabilizing the grid in critical situations. Proposing nuclear energy as part of a stable energy mix with highly fluctuating renewable sources overlooks its weaknesses. For technical and economic reasons, nuclear power plants do not provide the kind of flexible and programmable energy that can fill the gaps between solar production peaks.

In the face of hypothetical nuclear scenarios, it is therefore essential to look at the present: Italy is experiencing an unprecedented season in the field of renewables. In May 2025, renewable sources covered 55.9 percent of national electricity demand, a historic record, while photovoltaics recorded a 27-percent increase compared to the same month last year. The role of solar energy has also exploded in Europe: in June 2025, photovoltaics became the E.U.’s main source of energy for the first time, accounting for 22.1 percent of electricity production, surpassing gas, coal, and even nuclear energy. This signal is powerful and irrefutable: renewables are no longer a promise but a reality that is growing rapidly and strongly.

Terna, the transmission network operator, reminds us that fluctuating generation needs adequate storage systems as we continue to install solar and wind power. Short- and long-term capacity is needed; lithium-ion batteries do the day-to-day work, while pumping and Long-Duration Energy Storage (LDES) are building the future with pilots and commercial plants already in operation.

We cannot ignore this fact: Italy—and with it Europe—is demonstrating that a reliable energy transition can be based on renewables. When rethinking nuclear power, we need to place this reflection in the current context. SMRs could be a valuable complement if, and only if, they meet the demand for stability and do not shift it to uncertain terrain and timescales. We must embrace a coherent vision: efficiency, smart grids, storage, distributed generation, renewables, and—only where necessary—advanced nuclear power. Only in this way can we build a truly resilient, sustainable, and democratic electricity system.

Italy is now at a crossroads. As a country that has twice said “no” to nuclear power, it could choose to explore new solutions, such as SMRs, which are currently only a hypothesis with very uncertain timescales and costs. But at the same time, it is already undergoing another revolution: that of renewables and storage, which are already proving capable of covering more than half of our consumption and have taken the lead in Europe. As highlighted by IRENA3, in 2024, 91 percent of new renewable projects have been cheaper than fossil fuels, and 87 percent of new power capacity installed in the world has been renewable; in the past ten years lithium-ion batteries’ LCOS4 has fallen substantially by more than half. Between historical memory and the future, between atomic promises and solar certainties, the real challenge will not be to choose a single path but to compose a coherent mosaic of solutions, in which renewables are the backbone and nuclear power, if it ever really enters the picture, a complement.

For this reason, more than ever, transparency, stable rules, democratic participation and independent data are needed. The World Nuclear Industry Status Report remains the benchmark for distinguishing reality from illusion. It is from here that Italy will have to decide whether and how nuclear power can be part of its energy future. Only then will we be able to clearly assess whether nuclear power can be part of our real energy future, or whether it risks remaining yet another unfulfilled promise.

Key Insights

Executive Summary and Conclusions

The World Nuclear Industry Status Report 2025 (WNISR2025) provides a comprehensive overview of nuclear power plant data, including information on age, operation, production, construction and decommissioning of reactors. WNISR2025 includes various topical focus chapters. These include, for the first time, an in-depth assessment of the multiple Challenges of Integrating Nuclear Power into the Energy System as the system changes rapidly on all levels and in all geographical regions. The revamped traditional chapter Nuclear Power vs. Renewable Energy Deployment provides a complementary detailed overview of the astonishing expansion in the renewables sector, especially in combination with the soaring deployment of battery storage capacity on the one hand, and the stagnating nuclear sector on the other.

The Focus Countries chapter includes a detailed overview of developments in eight of the 31 nuclear countries including the top 5 generators (by rank)—i.e., the United States, China, France, Russia, and South Korea—as well as Japan, Ukraine, and the United Kingdom, plus Taiwan that closed its last reactor in May 2025. The chapter on Potential Newcomer Countries includes an Africa Focus section that assesses the status of planning in four selected countries on the continent. In addition, the chapter analyzes the status of nine other potential newcomer countries, including the only three (Bangladesh, Egypt, Türkiye) that are currently building their first commercial reactors. The traditional chapter on Small Modular Reactors (SMRs) reports many plans, particularly in potential newcomer countries, and finds increasing expenditures but still sees little progress on the ground.

Russia Nuclear Interdependencies follows up on the Russia Nuclear Dependencies chapter in WNISR2024 and looks into the mutual nature of dependencies between nuclear industries in the West and Russia. The Fukushima Status Report has been thoroughly upgraded, including a deep dive into the Japanese monitoring system intended to guarantee food safety. The Decommissioning Status Report now covers 218 closed reactors, almost one third of all the units ever connected to the grid over the past 70 years.

Finally, Annex 1 presents overviews of nuclear power programs in each of the countries not covered in the Focus Countries chapter.

Production and Role of Nuclear Power

Reactor Operation and Capacity. As of 1 July 2025, a total of 408 reactors—excluding Long-Term Outages (LTOs)—were operating in 31 countries, which is the same as the number of reactors reported in WNISR2024, but operating in one less country.5 There were 33 units in LTO, including 19 in Japan and 6 in Ukraine. Operating units are also 10 less than in 1989—the first peak and the end of the preceding uninterrupted rise—and 30 below the historic peak of 438 in 2002. As of the end of 2024, the world fleet had a total electric net operating capacity of a record 369.4 GW that had slightly declined by mid-2025 to 368.7 GW, which is just 1.6 GW—the equivalent of a large reactor—more than the previous end-of-year record of 367.1 GW in 2006.

IAEA versus WNISR Assessment. Between September 2022 and April 2023, the International Atomic Energy Agency (IAEA) significantly modified its statistics—including retroactively—in its online-Power Reactor Information System (PRIS). As of 1 July 2025, the IAEA-PRIS statistics show a peak of 440 reactors operating globally in 2005, whereas, as of the end of 2024, the operating capacity reached 377 GW, a new record, marginally over the previous peak of 374 GW reached in 2018, according to the new IAEA data. As of mid-2023, the IAEA had pulled, retroactively since shutdown, 23 units in Japan and four reactors in India from its list of operating reactors and added them to a new category labeled “Suspended Operation”. Following four subsequent restarts in Japan, as of mid-2025, the IAEA classified 19 reactors in Japan and four units in India as suspended.

As of 1 July 2025, WNISR classified 33 units as in LTO, of which 19 were in Japan, six in Ukraine, three in India, two each in Canada and South Korea, and one in China—the number decreased by one compared to WNISR2024.

Nuclear Electricity Generation. In 2024, the world nuclear fleet generated 2,677 net terawatt-hours (TWh or billion kilowatt-hours) of electricity. After a 4.4 percent drop in 2022, production increased by 2.2 percent in 2023 and again by 2.9 percent in 2024. It is the highest output ever, the previous record being 2,663 TWh in 2006. China, with a 3.7-percent increase (compared to 11.3 percent in 2021, 2.5 percent in 2022, and 4.1 percent in 2023), produced more nuclear electricity than France for the fifth year in a row and remains in second place—behind the U.S.—among the top nuclear power generators. Nuclear production outside China increased by 2.8 percent, similar to the level of global production in the mid-1990s prior to China’s nuclear buildout. If global nuclear generation exceeded the previous 18-year-old record by a modest 14 TWh—hardly more than a single large reactor’s nominal annual generation—outside China, nuclear generation in 2024 was 363 TWh below the 2006 level, a significant fall of nearly 14 percent.

Share in Electricity/Energy Mix. Nuclear energy’s share of global commercial gross electricity generation in 2024 was almost stable at 9 percent (–0.13 percentage points), the lowest value in four decades and more than 45 percent below the peak of 17.5 percent in 1996.

Reactor Startups and Closures6

Startups. In 2024, seven reactors were connected to the grid, three in China and one each in France, India, the UAE, and the U.S., leaving zero active construction projects in France, the UAE, and the U.S.

In the first half of 2025, one unit was connected to the grid in India—remarkably, not a single unit was connected in China or elsewhere.

Closures.7 Four units were closed over the year 2024, two in Canada and one each in Russia and Taiwan. In the first half of 2025, two were closed, one each in Belgium and Taiwan.

Over the past two decades (2005–2024), there were 104 startups and 101 closures in the world. Of these, 51 startups were in China, which did not close any reactors. As a result, outside China, the net number of reactors has significantly declined by 48 units and net capacity has declined by close to 27 GW over the period.

Construction Data8

As of 1 July 2025, 63 reactors (65 GW) were under construction, which is four more than in WNISR2024 but six fewer than in 2013; six of the 69 units listed in 2013 have subsequently been suspended or abandoned.

Eleven countries are building nuclear plants, two down from mid-2024 and five fewer than in WNISR2023. While Pakistan entered the category with the launch of a new construction at one of its existing sites, three countries dropped off the list: France completed its last construction project, Argentina abandoned a reactor that had been under construction since 2014, and Japan was taken off the list because the construction projects considered in earlier editions turned out to be inactive. Only four countries—China, India, Russia, and South Korea—have construction ongoing at more than one site. Three “Potential Newcomer Countries”—Bangladesh, Egypt, and Türkiye—are building their first nuclear power plants.

As of mid-2025, there is not a single active power reactor construction on the entire American continent, from Alaska to Cape Horn, including the U.S. that hosts the world’s largest nuclear power fleet.

Building vs. Vendor Countries

• As of mid-2025, China had by far the most reactors under construction with 32 units—five more than one year earlier and more than half of the total worldwide. In December 2024, China started work on its only active construction outside the country in Pakistan.
• Russia is by far the dominant supplier on the international market, with 27 units under construction as of mid-2025. Seven of these are domestic. The remaining 20 units are being constructed in seven countries, including four each in China, Egypt, India, and Türkiye.9 It remains uncertain to what extent these projects have been or will be impacted by the various rounds of sanctions imposed on Russia following its invasion of Ukraine. However, sanctions, including on the banking system, clearly have led to delays in various projects.
• Besides Russia’s Rosatom, only Électricité de France (EDF) and China National Nuclear Corporation (CNNC) are presently building abroad.

Construction Times

• For the 63 reactors being built, on average 5.3 years have passed since construction start—down from the mid-2024 average of 5.9 years—but many remain far from completion.
• All reactors under construction in at least six of the 11 countries hosting building sites have experienced, often year-long, delays.
• Of the 22–26 reactors behind schedule, at least 14 have reported increased delays.
• WNISR2023 noted a total of 14 reactors scheduled for startup in 2024, but only seven made it to generating first power, while the commissioning of the other seven was delayed at least into 2025.
• Grid connection of the Mochovce-4 reactor in Slovakia has been delayed yet again, currently to late 2025, that is 40+ years after its initial construction start. Bushehr-2 in Iran originally started construction in 1976, almost 50 years ago, and resumed construction in 2019 after a 40-year-long suspension. Grid connection was delayed again by one year and is currently scheduled for 2029, 53 years after construction initially started.
• Two additional reactors have been listed as “under construction” for a decade or more: the Prototype Fast Breeder Reactor (PFBR) and Rajasthan-8, both in India.

Construction Starts

• Construction started on nine reactors in 2024 (including six in China), up from six in 2023 and down from 10 in 2022 (both including five in China). Chinese companies also started one construction in Pakistan. Russia began work on one unit at home and one more in Egypt. In other words, over the five years between January 2020 and the end of 2024, all 40 construction starts in the world were either implemented by the Chinese or the Russian nuclear industries.
• Construction of five reactors started in the first half of 2025, three of them in China and one each in Russia and South Korea.

Operating Age

• The average age (from grid connection) of the 408 operating nuclear reactors has been increasing since 1984 and stands at 32.4 years as of mid-2025, up from 32 years in mid-2024.
• A total of 266 reactors, two-thirds of the world’s operating fleet, have operated for 31 or more years; of these, 141—more than one in three—have operated for at least 41 years.
• The age at closure remains remarkably low at 43.2 years on average for the 28 units closed in the five-year period from 2020 to 2024.
• The Plant Life Extension or PLEX Projection assesses how many reactors need to be built to compensate for expected closures. If all currently licensed lifetime extensions were maintained, all construction sites completed as planned, and all other units operated for a 40-year lifetime (unless an earlier or later closure date has been authorized), in the years to 2030, the net balance would be in the positive for 2026, drop into negative terrain in 2027–2029, followed by a sharp decrease in 2030. Overall, over the period 2025–2030, 44 new reactors or 26 GW in addition to the 59 units already under construction and scheduled to start up prior to the end of 2030 would have to be built and commissioned (or units restarted) to replace expected closures.
• The PLEX-Projection would still necessitate ramping up the annual startup rate of the past decade by a factor of 2.5 from 6.9 to 17.3 units for the remaining years to 2030 (including the reactors already expected to start up by 2030 (see Figure 21) only to maintain the status quo. However, probably at least half of the 104 reactors currently projected to close between 2025 and 2030 are now seeking and likely to secure a lifetime extension beyond 2030.

Focus Countries

The following nine Focus Countries include six of the top 10 nuclear power generators in the world. Some key developments for 2024 and the first half of 202510 include:

China. Nuclear power generation increased by 3.7 percent and provided 4.5 percent of total electricity generation, slightly slipping for the third year in a row. While nuclear capacity grew by 3.5 GW in 2024, solar capacity alone grew by 278 GW. Solar and wind together generated about four times more electricity than nuclear reactors. Since 2010, the output of solar increased by a factor of over 800, wind by a factor of 20, and nuclear by a factor of six.

France. Nuclear power generation increased by 13 percent, but at 362 TWh it remained significantly below the 400 TWh considered normal a decade ago. Nuclear power represented 67 percent of the country’s total power generation, the highest nuclear share in the world, but less than 18 percent of final energy. While total outage-days at zero-production dropped again in 2024 to, still significant, 99 days or a quarter of the year per reactor, the declared “forced” outages at 342 days stayed at the second highest level in six years. The Flamanville EPR started up in December 2024, twelve years later than planned, at a cost of US$25.6 billion, a staggering sixfold increase over the original cost estimate of US$4.3 billion (both figures in real 2023-dollars). The decision on the newbuild program has been confirmed; cost estimates are to be updated by year-end, and the first scheduled startup is delayed to 2038. Currently, there is no reactor under construction in the country.

Japan. Two additional reactors were restarted since mid-2024, bringing the total operational units to 14, while 19 reactors remain in LTO. Nuclear power generation increased by 9.5 percent, and the nuclear share in total electricity rose slightly from 7.7 percent in 2023 to 8.4 percent in 2024. Other restart schedules have been delayed again, and for the first time the Nuclear Regulation Authority (NRA) rejected a restart application on the grounds of non-compliance with regulatory safety standards. The operator of Tsuruga-2 could not prove there was no active seismic fault below the site. On the policy side, the goal to “reduce dependence on nuclear energy as much as possible” has been deleted from the Seventh Strategic Energy Plan, which sets out to “maximize the use of both renewables and nuclear power.”

Russia. State-controlled Rosatom is the leading global nuclear power plant exporter and the fourth largest nuclear electricity producer, contributing 17.8 percent of national power generation (down from 20.3 percent in 2020). Nuclear output has declined for the second and nuclear share in electricity for the fourth year in a row. For the third year in a row, Rosatom maintained its proactive role in the hostile military occupation of Europe’s largest nuclear power plant, Zaporizhzhia, in Ukraine.

South Korea. The country operates the fifth largest nuclear power program by capacity and production in the world. In May 2025, the South Korean nuclear industry became the first other than the Chinese and Russian (at home and abroad) to start construction since December 2019. The future of the nuclear industry remains uncertain nevertheless. Incoming President Lee Jae-myung is ambiguous about the expansion of nuclear power and favors a renewables-based strategy. This did not prevent the national nuclear utility—Korea Electric Power Corporation (KEPCO)/Korea Hydro & Nuclear Power Co. Ltd. (KHNP)—from signing a construction contract in the Czech Republic in spite of the group’s consolidated massive debt that, as of year-end 2024, stood at an unparalleled US$150 billion with revenues at US$69 billion.

Taiwan. With the country’s last reactor closing in May 2025, the country has completed its nuclear phaseout.11 Five other units had been closed previously in the framework of the strategy. Natural gas, contributing 42 percent, is now the largest source of electricity generation. So far, the buildup of renewables is lagging target levels, and the country remains dependent on imports for 95 percent of its overall primary energy supplies.

Ukraine. Of 15 operating or operable reactors, six are at the Russian occupied Zaporizhzhia site and stay in the LTO category as of mid-2025. The remaining operating reactors are a constant cause of concern in a country engaged in a full-blown war. Nevertheless, at over 50 percent, Ukraine has the third highest nuclear share in total electricity generation in the world. But production is down almost 40 percent compared to pre-war levels, as consumption also significantly shrank. Westinghouse is partnering with Ukrainian companies in a project to build two AP-1000s (of potentially nine in the country) at the Khmelnytskyi site. However, three years after the signature of a framework agreement, construction is yet to begin.

United Kingdom. There are only nine reactors with a combined capacity of 5.8 GW left operating versus 36 closed units in the country. Nuclear power provided a stable 38.6 TWh contributing 14.3 percent of the total supply (down a marginal 0.4 percentage points year-on-year but down from 28 percent in 1997). On 22 July 2025, after the WNISR mid-year editorial deadline, the U.K. Government signed the Final Investment Decision for the construction of two EPRs at the Sizewell C site. Meanwhile, in 2024, wind power alone provided more than twice as much energy as nuclear plants.

United States. Nuclear output increased slightly (+0.9 percent) to 782 TWh in 2024. The nuclear share of commercial electricity generation decreased by 0.4 percentage points to 18.2 percent. As of mid-2025, the U.S. nuclear fleet is still the largest in the world with 94 units. It is also one of the oldest with a mean age of 43.7 years. There is not a single nuclear reactor under construction in the country. There are many initiatives, plans, projects, financing schemes, and even presidential orders to restart building nuclear plants, big and small, but little is happening on the ground. A noteworthy development is Holtec successfully obtaining regulatory approval for moving the Palisades reactor in Michigan, officially closed in 2022, back into “operational status”.

Fukushima Status Report

Fourteen years have passed since the East Japan Great Earthquake on 11 March 2011 triggered the start of the Fukushima Daiichi nuclear power plant disaster (also referred to as 3/11 throughout the report). The situation is still far from having been stabilized.

Overview of Onsite and Offsite Challenges

Onsite Challenges

Fuel Debris Removal. In November 2024 and in April 2025, two samples of fuel debris were taken from the molten core of Unit 2. The total amount was 0.9 grams. There are an estimated 880 tons of fuel debris in Units 1–3, about a billion times the amount sampled, to be eventually extracted from the reactors and safely stored somewhere. There is no concept yet for a pathway to this goal, which is scheduled to be completed by 2051.

Spent-Fuel Removal. All spent fuel from the pool of Unit 3 had been removed by February 2021. Preparatory work is still underway on Units 1 and 2, with removal to begin in FY2026 for Unit 2 and in FY2027–2028 for Unit 1. TEPCO had completed spent-fuel removal from the pool of the relatively little impacted Unit 6 by mid-April 2025 and planned to start removal from Unit 5 in July 2025. TEPCO aims to have all the spent- and fresh-fuel assemblies removed from Units 1–6 by 2031, around 20 years after the disaster began.

Contaminated Water Management. As water injection continues to cool the fuel debris, highly contaminated water continues to run out of the cracked containments, mixing with water from an underground river in the basements. Various measures have reduced the influx of water from up to 540 m3/day in 2015 to about 50 m3/day in the first quarter of 2025. Nonetheless, an equivalent amount of water is partially decontaminated and stored in 1,000-m3 tanks daily, with a new tank filling up every three weeks.

The safety authority has allowed operator TEPCO to release treated contaminated water into the ocean, and as of the end of April 2025, TEPCO had discharged a total of 94,000 m3. As of the end of March 2025, two thirds of the 1.2 million m3 of stored water needed to be treated again, and all of the water has to be diluted by a factor of 100 or more in order to meet licensed standards before being released into the ocean. The ocean discharges remain contested in Japan and overseas.

Offsite Challenges

Offsite, the future of tens of thousands of evacuees, the contamination of food, and the management of decontamination wastes, all remain major challenges.

Evacuees. Although down from a peak of nearly 165,000 in May 2012, nearly 25,000 residents of Fukushima Prefecture remained living as evacuees as of 1 February 2025. The total number of people who have returned to former evacuation zones is uncertain and varies widely between locations. About 2.2 percent of the Fukushima Prefecture surface continues to be designated as “difficult-to-return zone”.

Disposal of Contaminated Soil. The soil, leaves, timber, and other waste removed in decontamination efforts is currently stored in interim facilities. There was around 14 million cubic meters of waste as of the end of December 2024, enough to fill 5,600 Olympic swimming pools, even though much of the heavily forested prefecture remains untouched. The government plans to use, e.g. in a range of construction applications, the soil that is judged safe for reuse, with radiation levels below 8,000 Bq per kilogram. The plan has met with significant opposition where trials have been proposed. The more heavily contaminated soil shall be shipped outside Fukushima Prefecture, but the destination and processing remain unclear.

Food Contamination. The section provides an analysis of the food monitoring system’s evolution over the years. Testing for radionuclides in most food stuffs has shown vanishingly few samples that exceed legal contamination limits. There are noteworthy exceptions though. In FY2024, 29 percent of the wild boar meat from Fukushima Prefecture that was tested exceeded radiation limits, with one specimen registering a cesium level of 13,000 Bq/kg, 130 times the legal limit of 100 Bq/kg. High levels in wild game have also been identified in other prefectures. Testing guidelines are vague, leaving details such as where and how much of various products are tested to the local governments, which implement them in different ways. There is no reliable, centralized data collection of testing results either. The interpretation of the significance of testing statistics is virtually impossible. The food monitoring system remains opaque and appears disorganized, making it challenging for the government to convince international observers, and even its own citizenry, that it has control over the situation. This has not stopped most of the 55 countries and regions that initially halted imports of food from Japan from lifting those bans, with only China, Russia, South Korea, Taiwan, Hong Kong, and Macau still restricting some products as of mid-2025.

Decommissioning Status Report

As an increasing number of nuclear facilities either reach the end of their operational lifetime or close due to excessive lifetime extension costs, timely decommissioning is becoming a key challenge.12

• The number of closed power reactors reached 218 units by mid-2025—five more than a year earlier and about one third of all reactors connected to the grid in the past 70 years. These had a total operating capacity of 110 GW. Over 100 units were closed in the past 20 years alone.
• 195 units are awaiting or are in various stages of decommissioning. Of these, 96 are in the “warm-up stage”, 33 in the “hot-zone stage”, 24 in the “ease-off” stage, and 42 in “long-term enclosure”.
• Only 23 units—about 10.5 percent of the closed reactors—have been fully decommissioned, no change from a year ago: 17 in the U.S., four in Germany, and one each in Japan and Spain. Of these, only nine, or 4 percent of all closed reactors, have been released from regulatory oversight as greenfield sites.
• The average duration of the decommissioning process is about 20 years, with a large range of 6–45 years (both extremes are for reactors with very low power ratings of 22 MW and 63 MW, respectively).
• To date, four of the early major nuclear states—Canada, France, Russia, and the U.K.—are yet to fully decommission a single reactor.

Potential Newcomer Countries

Potential Newcomer Countries with Active Reactor Construction

There are three potential newcomer countries with active reactor construction: Bangladesh, Egypt, and Türkiye. All the projects are implemented by the Russian nuclear industry.

Bangladesh. Two reactors of the Russian VVER design have been under construction since 2017–2018. They were scheduled to start up in 2023 and 2024. Sanctions have reportedly led to delays in the delivery of some equipment, and the commissioning of Unit 1 had been pushed back to December 2024, which also passed. In April/May 2025, employees of the local building company went on strike, followed by the dismissal of at least 18 engineers. The commissioning date of the plant is uncertain.

Egypt. Construction of the first nuclear power plant comprising four VVER-1200 reactors was launched at the El Dabaa site on 20 July 2022, even as the war in Ukraine was ongoing. Building of Units 2, 3, and 4 began in November 2022, May 2023, and January 2024, respectively. Startup of the first unit was previously planned for 2026 but has been pushed back to 2028.

Türkiye. Construction of the first of four VVERs started in April 2018, with Units 2–4 following in April 2020, March 2021, and July 2022, respectively. The first unit was to start operating in 2023 but that has been delayed several times. Current estimates are contradictory, putting commissioning at late 2025 or delayed again into 2026. There were multiple reasons for delays including technical issues on site, impacts of sanctions, and a wave of staff illnesses. In the latest development, as of mid-2025, reportedly 10,000 of 14,000 Russian employees returned home because they had not been paid.

Potential Newcomer Nuclear Countries in Africa

Only South Africa operates two aging reactors in continental Africa (see South Africa in Annex 1). China and especially Russia have been the most aggressive promoters of nuclear power on the continent. More recently the U.S. also started promoting nuclear technology on the continent, often marketing SMRs. These efforts focused on economically stronger countries such as Ghana and Kenya, with which the U.S. concluded nuclear agreements in the past year. China in turn is also involved in the implementation of large non-nuclear projects, solar in particular, on the continent.

Ghana. The country has set up a Nuclear Regulatory Authority, the Ghana Atomic Energy Commission with a Nuclear Power Institute and the company Nuclear Power Ghana to develop the first nuclear power plant project. The U.S. considers Ghana an important ally in the region, and a U.S.-Japan initiative aims at establishing Ghana as an African leader in SMR rollouts. While Ghana’s representative stated at the 2024 IAEA General Conference that “Corporation Framework Agreements” for an SMR and a large reactor had been signed, there are no official public documents and no company has communicated on the issue. Also, the country’s total installed power generating capacity of around 5 GW would not allow for the integration of a large reactor.

Kenya. The country has set up an official Nuclear Power Energy Agency (NuPEA) and plans to start construction of a large reactor in 2027, which in turn has amplified opposition from local communities and non-governmental organizations as well as parliamentary opposition. In a surprise move, in early 2025 the government announced its intention to dissolve NuPEA in the framework of a major restructuring of state institutions and relocate its functions elsewhere. Nuclear projects will not go far with the allocated budget of less than US$6 million for FY2025/2026.

Nigeria. The country signed nuclear cooperation agreements with several countries and considered the option of developing up to 4 GW of nuclear capacity. An official of the Nigerian Atomic Energy Commission presented in late 2024 a proposal to start building a nuclear plant by 2028/2029 with startup scheduled for 2034. Despite these pronounced ambitions, it is telling that the 2024 Nigeria Integrated Resource Plan, the country’s official roadmap for electricity generation, has not included any nuclear power in its projections to 2045. There is indeed no indication that Nigeria has in the past year moved any closer to implementing a nuclear power program.

Uganda. The country offers a striking illustration of the disconnect between reality and plans for nuclear development: the Ugandan government in May 2025 reaffirmed its intent to build 24 GW of nuclear capacity, 12 times the country’s total installed capacity as of mid-2024, and signed contracts for 26-month pre-feasibility studies with several Korean companies. How this could possibly fit with the timetable for the startup of a large reactor by 2031, as indicated by the minister in charge, remains a mystery.

Case Studies: Italy and Poland

Italy. The country was one of the first ones to start operating a nuclear power plant in 1963, Latina, which operated until November 1986. One year later, the Italian people, shocked by the Chornobyl disaster in 1986, decided in a national referendum to abandon the commercial use of nuclear energy. No reactor has generated power after the referendum, and Italy became the first of now five countries to phase out the use of an active commercial nuclear program. A second referendum was held in June 2011, only three months after the Fukushima disaster started unfolding. Then-President Berlusconi had planned to reintroduce nuclear power and had passed legislation to allow for nuclear newbuild, but 94 percent of Italian voters rejected the law.

Fourteen years later, Giorgia Meloni’s government contemplates the reintroduction of a nuclear program, while decommissioning of the facilities of the first program is still ongoing and there is no site for a final nuclear waste repository yet. The government has set up a National Platform for Sustainable Nuclear (PNNS) and its National Energy and Climate Plan (NECP) sees “great potential to develop new nuclear technologies for Italy.” A “with nuclear” scenario sees nuclear power covering around 11 percent of the electricity demand in 2050. Three major energy companies established Nuclitalia to assess market opportunities with an initial focus on SMRs. A Bank of Italy assessment warns about “the uncertainty surrounding the technologies chosen, most of which are not yet available for commercialization,” and calls for “a cautious approach that also prepares and promotes alternative strategies”.

Poland. The country already attempted to start a nuclear program; two reactors had been under construction for two years, when the Chornobyl disaster began in April 1986, which led to the abandonment of the project. Poland is one of a number of countries that started building their first reactors but abandoned prior to completion. There were several failed attempts to restart a nuclear program. The latest Polish Nuclear Power Program released for public consultation in early 2025 targets the deployment of 6–9 GW of nuclear capacity with somewhat unclear timelines. Original construction start dates were delayed, and as of December 2024, first concrete for the first unit was scheduled for 2028 with the aim of consecutive grid connections of the first three Westinghouse AP-1000 units with up to 3.75 GW between 2036 and 2038. Westinghouse has partnered up with U.S. construction giant Bechtel and the Polish state-owned project company Polskie Elektrownie Jądrowe (PEJ). Cost estimates have doubled since 2022 to over US$48 billion. A second large-reactor project led by Korean KHNP and various SMR designs are also being discussed but remain in earlier stages.

Other Examples

An increasing number of countries are declaring plans to introduce nuclear power programs. Many of these claims either lack credibility or imply timeframes that are at this point irrelevant for an annual report on the industry. Early potential newcomers include Ecuador, Estonia, Indonesia, Jordan, Kazakhstan, Saudi Arabia, and Uzbekistan that are briefly analyzed hereunder.

Ecuador. In the fall of 2024, a senior member of government had unveiled an “ambitious roadmap” to deploy a 300-MW reactor by 2029, with a 1-GW reactor to follow in the longer term. No official document is available. The country has no regulatory framework and no institutions like a nuclear regulator or a nuclear waste management agency, and it is very uncertain whether the national grid could deal with a nuclear plant of any size.

Estonia. In 2023, an interministerial Nuclear Energy Working Group (NEWG) concluded that the introduction of nuclear energy in Estonia was feasible and that SMRs with a capacity of less than 400 MW would be suitable for the country. While a resolution supporting the adoption of nuclear energy allowing for the drafting of appropriate legislation was adopted in parliament in June 2024, it saw 27 of 68 MPs voting against (25) or abstaining (2). This was an unexpected political signal in the very early stages of the potential introduction of a nuclear program.

Indonesia. Indonesian government representatives have long been talking about nuclear power plans. More recently, a senior official told the media that Indonesia was planning to operate 10 GW of nuclear capacity by 2040 and that companies from various countries “have shown interest”. The country is looking for IAEA assistance to develop a comprehensive roadmap towards a nuclear program but seems a long way from actually building reactors.

Jordan. Set up in 2008, the Jordan Atomic Energy Commission developed various plans for the construction of large nuclear reactors in what is one of the water poorest countries in the world. Ten years later, these plans were abandoned, and the focus has since been redirected to SMRs. Jordan has so far not selected a design, vendor, site, or completed a financing package, let alone started building any nuclear power plant, small or large.

Kazakhstan. The country is one of five to have discontinued the use of nuclear power in the past. It has been the world’s leading uranium producer for more than a decade. The restart of a commercial nuclear power program has been contemplated for the past two decades. In a national referendum held in October 2024 about 71 percent of the votes expressed support for a nuclear program. In the months before, civil society movements staged various forms of protest against the plan. In March 2025, the President created the national Atomic Energy Agency by decree. In June 2025, the agency announced Rosatom as the supplier of the first new Kazakhstani nuclear plant and a forthcoming agreement with China for a second plant.

Saudi Arabia. The King Abdullah City for Atomic and Renewable Energy (KA-CARE) was established in 2010. In September 2024, a government representative stated that “the Kingdom is moving towards utilizing nuclear energy.” But not much progress has been made over the 15 years since the establishment of KA-CARE. In the meantime, over the past decade alone renewable energy capacity has grown by a factor of 200, from 24 MW to 4.7 GW in 2024 (90 percent solar), yet it still represents only 2.2 percent of power generation.

Uzbekistan. In May 2022, officials announced that a site for the construction of two Russian-designed VVER-1200 reactors had been chosen. Subsequently, the plan was apparently abandoned in favor of an SMR project. Reportedly, in May 2024 the government signed an agreement with Russia’s Rosatom to build six 55-MW Small Modular Reactors (SMRs) in the eastern Jizzakh region. In April 2025, Rosatom announced that onsite work had begun to set up production facilities, administrative buildings, and warehouses for the SMR cluster. By June 2025, the project had circled back to include large units.

Russia Nuclear Interdependencies

Russia is a major global supplier of nuclear fuel services, including uranium mining, conversion, enrichment, and fuel assembly fabrication for Soviet-designed VVER pressurized water reactors, of which there are 19 in the E.U. and 15 in Ukraine. Since Russia’s full-scale invasion of Ukraine in February 2022, E.U. members and others in the region have discussed and taken measures to deprive Russia of the considerable revenue streams provided by such businesses and reduce the inherent risk of the region’s dependence on Russia. While the U.S. banned the import of uranium products from Russia in May 2024—to which Russia responded in November 2024 with a tit-for-tat restriction on enriched uranium exports—the E.U. has not established any sanctions in the nuclear sector, a strong indicator of its dependency on Russia. However, the dependency is multifaceted; Russia is also dependent on the West for European companies’ cutting-edge technology used in its reactors, and business interests of these same companies have certainly contributed to the prevention or delays of sanctions.

In 2024, U.S. imports of enriched uranium from Russia dropped by about half compared to 2023, while imports from China grew to the sixth largest in 2023–2024 after zero imports in 2020–2022. In the E.U., Rosatom’s share decreased in 2024 after a spike in 2023 in the three categories natural uranium, uranium conversion, and uranium enrichment, but it still provided 16–24 percent of the respective services in 2024 compared to a 23.5–38 percent share in 2023. Imports of fuel assemblies for 17 Soviet-designed VVERs in four E.U. countries—Czechia, Finland, Hungary, and Slovakia (Bulgaria with its two units is not included)—have also declined in 2024 to 438 tons from a peak of 573 tons in 2023, though they still exceed the 314 tons in 2022. Four of the five countries, except Hungary, have signed supply contracts with Westinghouse, the only operational alternative, so far, to Rosatom subsidiary TVEL for VVER fuel. In addition, four of the five, excluding Finland, have signed contracts with EDF subsidiary Framatome for VVER fuel. However, Framatome does not yet have any operational manufacturing plant. A Lyon-based Russian-French joint venture chose Framatome’s Lingen site in Germany for a VVER-dedicated fuel-assembly production-line project that ran into political problems with the German authorities and significant public opposition that, one year later, still remain to be resolved.

The long-standing business relations between Framatome, its partner Siemens Energy, and Rosatom create a mutual dependency, including significant Russian dependence on the West which may get increasingly relevant when further sanctions are considered. For at least a decade Russia has received from Framatome and Siemens Energy Instrumentation & Control (I&C) technology not only for the initial construction but also modernization and servicing of existing reactors which last long into the future. The most recent reported case involves a Framatome I&C system, due to be fully installed at the newbuild Kursk II nuclear power plant (two units under construction) by the end of 2025.

With the exception of one Chinese project in Pakistan and one “first concrete” in South Korea, Rosatom implemented all 16 nuclear power reactor construction sites started outside China over the past five years and a half, and providers of parts, e.g., France’s Arabelle turbines, do not have any other foreign customer besides Rosatom.

Small Modular Reactors (SMRS)

The first time WNISR published a chapter dedicated to SMRs it read: “One reaction to the decline of nuclear power from the nuclear industry and other proponents has been to advocate what it terms ‘advanced reactors’. (…) In the last decade, the overwhelming focus of this effort has been on what are called Small Modular Reactors (SMRs).” That was in 2015, a decade ago. The conclusion then was: “For decades, small modular reactors have been held out as holding great promise for expanding nuclear power into various new markets. Remarkably similar claims have been made by the nuclear industry in multiple countries. Nowhere have these claims come true.” Today, the conclusion is very similar, except that the gap between hype and industrial reality has never stopped growing.

A significant development is that multiple governments are now spending large amounts on SMRs, and international organizations and financial institutions are starting to make financing available. Some startups are also raising increasing amounts of private money, even if generally only complementing generous public matching grants. The OECD’s Nuclear Energy Agency (NEA) has estimated the total public/private funding made available worldwide for SMRs at US$15.4 billion. The NEA included 74 SMR designs in its SMR dashboard of a total of 127 it has counted, which means the amount would be thinly spread, unless a large number of designs are abandoned. Just one company, NuScale, has already invested around US$2 billion and has still not started any construction anywhere.

Country-by-Country Status Overview

Argentina. The CAREM-25 project had been under construction since 2014. Construction was halted in spring 2024 due to budget cuts. By September 2024, 470 workers had been laid off, and in December 2024, the president of the National Atomic Energy Commission (CNEA) declared the end of the project as “this reactor is not economically competitive.” Another SMR design also under development in Argentina called ACR-300 is now the focus of attention. A former CNEA president commented that “this kind of announcement of the Argentine government is to say something for people [to] cheer” because the ACR-300 “has no engineering detail of any kind.”

Canada. In April 2025, the Canadian Nuclear Safety Commission (CNSC) approved a license request by Ontario Power Generation (OPG) for the construction of a BWRX-300 reactor designed by GE-Hitachi at the Darlington site. This is significant as it represents the first SMR construction license in the Western world for a design that has not yet been built anywhere. Grid connection is scheduled for 2030. In May 2025, the Ontario Government approved OPG’s plan to spend US$5.6 billion on stage one of the project that shall ultimately include the construction of three more BWRX-300 units with a total investment of US$15.2 billion. Various other SMR projects in early stages are also under consideration in other provinces.

China. The country is operating or building two kinds of SMRs, a high-temperature gas-cooled reactor design called the HTR-PM and an integral pressurized water reactor design, the ACP100 (or Linglong One). There is little to no information available as to the operational experiences with the two HTR-PM modules. The nominal capacity has been reduced by one quarter from combined 200 MW to 150 MW for unknown reasons. No reliable information is available on its performance. The second design, called ACP100, has been under construction since July 2021 with scheduled startup by May 2026.

France. In mid-2024, EDF announced that it had suspended and reoriented the development of its project called Nuward towards “a design based on proven technological building blocks.” Soon after the announcement, TechniAtome, France’s nuclear submarine reactor maker, pulled out of the project consortium. In January 2025, the EDF subsidiary Nuward revealed that the new design would be a 400-MW reactor and offer an option for cogeneration (up to about 100 MWt). The power rating would exceed the limit of 300 MW usually attributed to the SMR definition, and the concept is in its very early pre-basic-design phase.13

India. The Indian government in February 2025 announced that there would be “at least five indigenously developed SMRs” in operation by 2033. There have also been announcements about imported SMRs. Whether small or large reactors, India’s plans to build or import nuclear technology have mostly not materialized. In addition, the many challenges confronting SMRs, especially their lack of economic viability, render the ambitious construction plans quite unrealistic.

Russia. Russia is building both light water and fast neutron SMRs, with a special focus on barge-mounted reactors for coastal locations. Two 30-MW “floating reactors”, the Akademik Lomonosov, were started up in December 2019 and have been underperforming since. At least two more barge-mounted projects are underway. Construction on a different, land-based SMR project, a lead-cooled fast reactor design called BREST-300, started in June 2021. The projected startup date has been delayed to 2028.

South Korea. In 2012, the System-Integrated Modular Advanced Reactor (SMART), a PWR design, received the safety authority’s approval, but did not find any buyers. Another design called the i-SMR is in early stages of development. The regulator has yet to receive an application for standard design approval, currently not expected before 2028, with plans to start construction delayed to 2031.

United Kingdom. Since 2014, Rolls-Royce has been developing the “UK SMR”, a (now) 470-MW reactor (exceeding the size-limit of 300 MW for the generally adopted SMR definition). The regulator is currently carrying out a Generic Design Assessment (GDA) that is scheduled to be completed by December 2026. Two other SMR designs are under GDA: Holtec’s SMR-160 and GE-Hitachi’s BWRX-300. Westinghouse and EDF had withdrawn from the competition. The U.K. Government has pledged a significant US$3.4 billion for the overall SMR program aiming at the commissioning of three units in the mid-2030s.

United States. The Department of Energy (DOE) is the largest funder of SMR development in the U.S. and awarded a combined US$2.8 billion in matching grants to X-energy, TerraPower, and Kairos Power. In November 2024, Kairos Power received a permit to build a facility with two 35-MWt test reactors called Hermes-2. While this permit is explicitly for a non-power facility (and thus excluded from WNISR statistics), theoretically the reactors could later be transformed into a power plant, upon approval of necessary license amendments. In May 2025, the Tennessee Valley Authority (TVA) submitted a permit application to the regulator to build a GE-Hitachi BWRX-300 SMR. In March 2024, TerraPower had submitted its application to build a 345-MW fast reactor called Natrium—exceeding the size-limit of the SMR-definition—in Wyoming. However, only one design, NuScale, has received a Standard Design Approval from the regulator but is not yet under construction anywhere in the country.

Challenges of Integrating Nuclear Power into the Energy System

This topical focus chapter examines the systemic implications of the fundamental differences between nuclear fission and renewable energy technologies and their respective roles within the overall energy supply and consumption systems.

New Energy Technologies Disrupt Markets and Systems

The declines in unsubsidized costs observed in photovoltaics, batteries, and power electronics over the past decade are unparalleled in the history of energy technologies. This has led to a spectacular acceleration of deployment. However, the new energy technologies are not merely replacing the old ones, they are fundamentally changing the logic of the system. As electricity becomes the central energy carrier, its generation becomes increasingly decentralized, and fluctuating wind and solar require the development of new flexibility options.

Nuclear Power Plants Need to Be Large

Due to several physical laws, thermal and mechanical devices are intrinsically more efficient, e.g., for power generation, when they are larger. These basic characteristics as well as the constant search for economies of scale have led to ever increasing average sizes of nuclear power plants. They also make the search for an economical small reactor rather illusionary. The idea of replacing economies of scale by economies of series production cannot compete with the advantages of mass production in the renewable energy sector. Over the past 20 years, an average of five reactors per year started up. If it was possible to increase the construction frequency by a factor of 10 or even 100, the series would remain very small in terms of industrial mass production. For comparison, a modern solar factory can produce two billion photovoltaic cells per year.

Photovoltaics—Essentially Electronic and Extremely Scalable

The real price of solar modules declined by 99.6 percent between 1976 and 2019. The conversion efficiency from solar radiation to electricity is currently around 20 percent, and next mass-market technologies already reach up to 35 percent in the lab. Gains in efficiency caused by advances in nanoscience-based material research are the main driver of solar electricity cost reductions. While production costs per module area may not further decrease substantially, efficiency is expected to continue to improve considerably. Solar cells have around 5–6 Watts each and can be assembled to modules of various sizes (medium size about 1.8 m2 and 400–700 Watts) and any number of modules can make electricity generating facilities of any size.

Underestimated Power Electronics

Wind turbines, electric cars, photovoltaic devices, and efficient electric motors would not be possible or would have a hard time competing against traditional systems without modern, semiconductor-based power electronics. Thanks to an improving understanding of the processes involved at the nanoscale and the development of new materials, the power density of power electronics devices has increased by a factor of 1,000 over the past 25 years. Power electronics are at the core of the upcoming digitally controlled network that allows for highly efficient power flows in all directions, flexibly adapting to the demands and offers of millions of power producers and consumers.

Wind Power’s Breakthrough with Power Electronics

Since the late 1990s, advancements in power electronics have enabled the complete decoupling of generator speed from the frequency of the power delivered to the grid. Furthermore, digital controls facilitate the continuous optimization of the blade pitch, thereby ensuring maximum rotor efficiency in variable wind conditions.

Renewable Power Generation—Unbeatably Cheap

Two recent authoritative, comparative assessments of the Levelized Costs of Energy (LCOE), i.e., the costs per kilowatt-hour over the lifetime of a facility, show a clear advantage of solar and wind power generation over traditional competitors. In a detailed study for Germany, the Fraunhofer Institute concluded that in 2024, utility-scale PV plants in southern Germany could produce at US$c4.5/kWh, whereas small rooftop installations, delivering at the retail level, at US$c6.8/kWh. Compared at the retail level, PV rooftop electricity costs were lower than the costs of PV utility scale if distribution costs were added. Onshore wind power came in at US$c4.7/kWh for good locations. The lowest estimates for natural gas were US$c9.6/kWh, for lignite US$c16.4/kWh, for hard coal US$c18.8/kWh, and for nuclear US$c14.8–53/kWh. In the U.S., according to Lazard, LCOEs for onshore wind are at US$c3.7/kWh, for PV utility US$c3.8/kWh, for offshore wind US$c7/kWh, and for PV residential and commercial US$c8.1/kWh, while the lower end for U.S. nuclear was at US$c14.1/kWh (3.7 times as much as utility solar) and for gas peakers US$c14.9/kWh. With US$c3.4/kWh, purely operational costs for nuclear power plants are in the range of onshore wind and utility PV.

Battery Costs Reach a Tipping Point for the System

Retail prices for ready-to-use home storage kits now start at US$200/kWh. Assuming levelized costs of storage between US$c3.4/kWh and US$c8.9/kWh (based on actual offers on the market) and that 100 percent of the electricity generated by rooftop PV is being stored to be delivered at any time during the following 24 hours, taking Fraunhofer’s lower values for rooftop PV in southern Germany quoted earlier and applying the two indicated levelized cost estimates leads to an overall cost for solar power delivered around the clock of US$c9.7–15.2/kWh—the latter corresponding to about half of the present utility power retail price in southern Germany. In reality, the storage need is of course much lower than 100 percent, which makes the option even more compelling. In China, a December-2024 auction reaching a price level of US$66/kWh of storage capacity has demonstrated what this means on the large scale: storing one kWh for about one cent.

The price decline has led to a surge of stationary storage capacity, growing with a compounded annual growth rate of 58 percent in Europe between 2022 and 2024. However, the vast majority of batteries, about 85 percent, are being used in cars. As most cars are sitting idle for over 90 percent of the time, the electric vehicle fleet has an extraordinary potential if bidirectional charging becomes the norm. If all cars in Germany were electrified, their batteries could power the country for two days of current consumption.

Nuclear Power and the Imperatives of the Energy System

Nuclear Power and the Climate Emergency. Nuclear power’s much longer lead times and far higher costs per MWh, compared to modern renewables, mean new nuclear plants will save less emissions per dollar as well as per year (see Climate Change and Nuclear Power in WNISR2019).

Lifetime Extensions. Upgrading and uprating existing reactors is possible but expensive and limited in scope. The output of a nuclear reactor can be increased by little more than 20 percent while repowering of wind turbines frequently doubles or triples their output. Repowering solar plants to double output is becoming attractive as well.

New Nuclear Reactor Designs. New concepts cannot eliminate the basic cost drivers: safety and security issues linked to the elementary nuclear forces involved and the inherent drive for size in thermal power plants, whatever primary energy they operate on.

Nuclear as a Dispatchable Power Source. For technical and economic reasons, nuclear plants do not provide the type of flexible, dispatchable power that can fill the gaps between solar and wind power peaks. They also need flexibility from other sources to bridge considerable planned and unplanned outages and for buffering services between changing demand and their inflexible full-load operation.

Nuclear Power and Data Centers. Building new nuclear plants for data centers appears incoherent. Time horizons do not match: while data centers need power in the short term, nuclear power plants need many years to develop, plan, and build; competing solar power plants can be set up within months (see also Nuclear Power vs. Renewable Energy Deployment).

Nuclear Power vs. Renewable Energy Deployment

A mix of contradictory trends has marked the year 2024. On the one hand, in most regions, inflation, rising interest rates, growing political uncertainties, regulatory stagnation, and declining investor confidence provided a more difficult environment for the energy transition. On the other hand, declining prices and advancements in key technologies have been encouraging and have opened new perspectives. In the first months of 2025, these trends became even more pronounced. For example, in April 2025, global solar electricity generation exceeded monthly nuclear power generation for the first time and kept doing so in May and June 2025.

Investment. For two decades, global investments in renewable power generation have exceeded those in nuclear energy—which hardly changed over the past decade—and are now 21 times higher. In 2024, solar PV saw the strongest growth at 22 percent compared to 2023. Wind power investments have shown lesser resistance to an accumulation of issues, including interest rate increases, inflation, supply chain pressures, as well as regulatory and political uncertainties, and have declined by 16 percent over the same year. Total investment in non-hydro renewable electricity capacity in 2024 was estimated at US$728 billion, up eight percent compared to the previous year. Most remarkable, however, is the 2024-surge in minor markets: Asia-Pacific, excluding China and India, +81 percent; Central and Southern Africa plus Europe, excluding the E.U. and U.K., +85 percent; and the Americas, excluding the U.S. and Brazil, +167 percent.

Installed Capacity. Solar and wind capacity grew by 452 or 32 percent and 113 GW or 11 percent, respectively, with grid scale battery capacity jumping 113 percent to 126 GW, according to the Energy Institute. These numbers compare with a net addition of 5.4 GW nuclear capacity between new startups and closures.

Electricity Generation. In 2021, the combined output of solar and wind plants surpassed nuclear power generation for the first time. In 2024, wind and solar facilities generated over 70 percent more electricity than nuclear plants. The evolution of the global electricity supply system since 2000 shows that coal’s part peaked in 2007 at 41 percent and declined to 34 percent in 2024, while renewables’ share (including hydro) increased from 19.4 percent to 31.6 percent between 2010 and 2024 alone. Over the same period, nuclear contribution fell from almost 13 percent to 9 percent.

In 2024, wind power generation grew by 8 percent or 188 TWh to 2,486 TWh—getting close to nuclear’s new record generation of 2,677 TWh—and solar output grew by 28 percent or 461 TWh to 2,091 TWh net. Total non-hydro renewable power generation increased by 670 TWh, which—and that is the bad news—covered only 52 percent of the global increase in electricity production of close to 1,300 TWh.

Nuclear power generation grew by 3 percent or 80 TWh. The new global record is marginally above the previous 18-year-old record (see above) but remains almost 14 percent below the 2006-generation level outside China, a significant decline.

Negative Power Prices and Batteries. The rapidly increasing deployment of stationary Battery Energy Storage Systems (BESS) might be a gamechanger for the increasing lack of flexibility in the grid systems. Negative power prices indicate temporary overproduction and illustrate the system’s difficulties in coping with fluctuating renewables. The number of negative hours across all power markets in continental Europe shot up by a factor of 10 between 2022 and 2024.

The acceleration of grid-scale BESS installations is impressive. Over 80 percent of the overall BESS capacity has been added in the past three years. The average annual global growth rate jumped to 115 percent between 2021 and 2024, with large geographical variations, achieving a total BESS capacity of 360 GWh. The grid integration of 350 GWh “batteries-on-wheels” in the U.S. alone would almost double global capacity and illustrates their huge potential.

Boom Behind the Meter. Decentralized PV installations represent significant shares of solar installations. On the global average, 42 percent of total solar capacity added in 2024 were rooftop systems. China, which had pushed the rooftop market share to 58 percent in 2022, scaled back to 44 percent in 2024. With the dramatic decline in costs of batteries, an increasing number of stakeholders—households, businesses, services, farmers—decide to add storage to their solar system. Many, especially in some Asian and African countries, are not connected to the grid. This has created some statistical uncertainty about the real scope of the ongoing transformation.

China. In 2024, China alone accounted for 40 percent of the global solar electricity generation, and the same is true for wind. But solar grew by 44 percent, more than three times as fast as wind which increased 13 percent. In contrast, China’s nuclear output accounted for 16 percent of global nuclear electricity production and grew only by 3.7 percent. As other sources have been growing so much faster, the nuclear share in power generation dropped slightly for the third year in a row to 4.5 percent in 2024, when solar and wind alone generated over four times more power than nuclear. Renewables, including hydro, increased from 18.7 percent in 2010 to 33.7 percent in 2024, while coal peaked in 2007 at 81 percent and declined to 57.8 percent in 2024.

European Union. In 2024, wind contributed 485 TWh or 17 percent to the overall electricity generation, while solar contributed 300 TWh or 11 percent. Fossil fuel generation has dropped to 793 TWh, hardly more than wind and solar combined. The wind plus solar contribution of 28 percent to the total electricity production, not only outcompetes nuclear’s 23-percent share in the E.U., but is also much higher than in China (18 percent) or the U.S. (17 percent) and nearly twice the global average (15 percent). The E.U. nuclear output increased slightly year-on-year (+28.5 TWh) due to the partial recovery of the French nuclear fleet (+41.3 TWh). In other words, outside France nuclear power generation declined by 13 TWh.

India. Solar capacity grew by 24 GW (33 percent) and wind by 3 GW, while solar power generation increased by 20 TWh and wind generation stagnated. A new reactor with 0.63 GW was connected to the grid early in the year, and nuclear generation increased by 6 TWh to a total of 52 TWh. Solar generated 135 TWh, 2.6 times the nuclear output, and wind produced 81 TWh, which is 56 percent more than the nuclear fleet.

United States. Wind and solar in 2024 have grown, respectively, by 3.5 percent and 28 percent in capacity and by close to 8 percent and 27 percent in electricity generation. Combined, wind with 453 TWh and solar with 303 TWh, are approaching nuclear power’s rather stable output of 782 TWh, the same level as already achieved in 2005. In recent years, Texas, a conservative, Republican-run oil state, has become the epicenter of stationary battery growth. By the end of 2024, Texas had an installed capacity of almost 10 GW of grid-connected batteries, expected to double in 2025 to 20 GW. Only three years earlier, in 2022, battery capacity stood at 2.8 GW. At the same time, Texas has been building up around 40 GW of solar and wind each and retiring fossil-fueled plants; notably, 7.3 GW of coal and gas closed between 2018 and 2023. It remains uncertain to what extent the current Trump administration’s pro-fossil fuels and pro-nuclear policies will impact the energy transition efforts on state level.

Introduction

The World Nuclear Industry Status Report 2025 (WNISR2025), more than ever, documents the widening and deepening gap between public perception and reality of the status and trends of the nuclear industry. The number of operating reactors is identical to a year ago as described in WNISR2024. Following the completed nuclear phaseout in Taiwan, there are less countries operating nuclear power plants. With France completing its last reactor under construction (third largest fleet in the world), Argentina abandoning a construction project it started in 2014, and Pakistan being the only additional country starting construction, there are fewer countries in total building power reactors. The average age of the world’s fleet keeps increasing.

Several governments announced major policy changes, financial commitments, and legal or regulatory changes in favor of operational lifetime extensions and nuclear newbuild. In some cases, public matching funds now attract significant private capital into design development. Remarkably, in some countries, there are initiatives to restart reactors that had been formally closed. The first case where operational status of a closed unit has been reinstated by a national regulator (in July 2025) was the Palisades reactor in Michigan, United States. Similar restart-from-closure initiatives in the U.S. involve Three Mile Island-1—the twin reactor of TMI-2 that experienced a partial meltdown in 1979—and Duane Arnold-1.

None of these initiatives, so far, have translated into trend changes. The only end-of-year 2024 or mid-2025 indicators that went up are entirely due to the ‘China effect’. The total number of reactors under construction slightly increased due to new projects getting under way in China. The operating capacity and with it the nuclear power generation reached a new record high of 2,677 TWh in 2024, a marginal 14 TWh (0.5 percent) above the previous record in 2006. In fact, compared to 2006, nuclear power generation outside China in 2024 was lower by 363 TWh (14 percent)—the equivalent of the combined 2024 output of the 21 smallest of the then-32 nuclear countries. Even in China, nuclear newbuild is dwarfed by the phenomenal expansion of renewables, solar in particular. Reportedly, China installed about 100 solar panels every second to add 93 GW of solar capacity in the single month of May 2025.14 Over the entire year 2024, China started up just three new nuclear reactors with a combined capacity of 3.5 GW; solar panels generated almost twice as much electricity as the country’s nuclear fleet, the second largest in the world.

The revamped traditional chapter Nuclear Power vs. Renewable Energy Deployment provides a detailed overview of the astonishing dynamic in the renewables sector, especially in combination with the soaring deployment of battery storage capacity and in comparison with the stagnating nuclear sector.

Eleven years ago, WNISR launched a first, modest attempt to look at system issues linked to the different characteristics of various power generating technologies (see Paying to Produce in WNISR2014). The analysis looked in particular at the issue of baseload power versus increasingly fluctuating production and economic implications of this. Ever since, the WNISR coordinator tried to design, and find a qualified author for, an in-depth assessment of the multi-layered, complex, increasing challenges of integrating nuclear power into a rapidly, drastically changing power system environment. The WNISR2025 offers just that, a 52-page assessment of the multiple Challenges of Integrating Nuclear Power into the Energy System as it emerges locally, regionally, nationally, globally. This chapter explains why renewables have turned into the undisputed drivers of the ongoing energy revolution. It has to do with physics, timescales, and economics. As Josephine Steppat, energy analyst at Montel Analytics, recently stated in a column on Montel News:

What is less understood is this is not just a policy debate but a clash of system logic. This makes nuclear power and renewables fundamentally incompatible as sources of energy in a modern grid. (…) Pursuing both new nuclear baseload and volatile renewables is not a coherent strategy – it is a conflict. The results: overcapacity, price collapses during renewables peaks, rising electric grid costs and inefficient investment on both sides.15 

Updated and expanded country focus chapters on China, Japan, Russia, South Korea, Ukraine and the United States look into some of the major developments over the year. The chapter Russia Nuclear Interdependencies follows up on the Russia Nuclear Dependencies chapter in WNISR2024. The slight modification of the title is not a coincidence. The chapter documents why the dependency of Western nuclear companies on Russia as a client is (at least) as significant as their dependency on Russia as a supplier of materials and services. After all, the international market for nuclear reactors is a niche seller’s market entirely dominated by Russia: all 13 construction starts outside China and Pakistan (where Chinese companies started building a reactor in late 2024) in the five years between 2020 and end of 2024 were implemented by Rosatom and its subsidiaries. If Russia was not buying the Western industries’ equipment and components, who would?

The traditional chapter on Small Modular Reactors (SMRs), in spite of overwhelming attention by policy makers and the media as well as a flurry of non-binding agreements, has little concrete progress to report. And while one construction permit has been issued in Canada, there is still not a single SMR under construction in the Western world (outside China and Russia). In fact, one SMR construction, in Argentina, has been abandoned in 2024, 10 years into the works. This makes the entire American continent, from Alaska to Cape Horn, currently nuclear construction free.

Many countries have signed nuclear cooperation agreements and have announced intentions to start national nuclear power programs. A selection of these initiatives can be found in the Potential Newcomer Countries chapter. Only three potential newcomers are in the course of building reactors (Bangladesh, Egypt, Türkiye); many plans fail the industrial and economic credibility test.

The Fukushima Status Report has been thoroughly upgraded, including a deep dive into the Japanese system that is supposed to guarantee food safety. The outcome is a description of a rather surprisingly disorganized system that makes it difficult or impossible to gain full independent understanding of the pertinence of statistics and their analysis made available by the Japanese government and other stakeholders.

The Decommissioning Status Report now covers 218 closed reactors, almost one third of all the units ever connected to the grid over the past 70 years. From mid-2024 to mid-2025, five reactors closed and four started up. As of end of August 2025, two reactors had closed (Belgium and Taiwan), but only one reactor had started up since the beginning of the year (India). Only one in ten closed reactors (23) have been technically fully decommissioned, while just nine, or one in 24, of the sites have been returned to “greenfield” status, free for any use.

 

Editor’s notes: The WNISR’s general editorial deadline is mid-year. In some cases, when authors or editors catch information that is post-1 July 2025 but considered essential or contradicting other information in the text and time permitting, it has been added to the body of the text or a footnote with the update has been inserted.

Throughout the report, to the best of our ability, amounts in foreign currencies that were allocated, estimated or spent during the ongoing year are translated into US$2025 by applying the monthly exchange rate as of June 2025—unless the amount is explicitly based on the value of a different year, in which case the yearly exchange rate of the corresponding year is applied.

Amounts in foreign currencies that were allocated, estimated or spent in past years are converted to the corresponding value by applying the yearly average exchange rate.

General Overview Worldwide

Role of Nuclear Power

The role of nuclear power in the global energy landscape is shrinking. As of mid-2025, there are fewer countries operating nuclear power reactors and fewer countries building reactors than one year earlier.

As of mid-2025, 31 countries operated nuclear power programs in the world. Figure 1 illustrates how the spread of nuclear power throughout the world took place at a significantly slower pace and smaller scope than anticipated in the early 1970s:

• Fourteen countries had operating nuclear power reactors (grid connected) when the Treaty on the Non-Proliferation of Nuclear Weapons (commonly known as the nuclear Non-Proliferation Treaty, or NPT) entered into force in 1970.
• Sixteen additional countries were operating power plants by 1985, the year when reactor startups peaked.
• Five countries (China, Romania, Iran, United Arab Emirates, and Belarus) started up power reactors for the first time over the past 30 years, of which two in 2020.
• The number of countries operating power reactors in 1996–1997 reached 32. It took another 23 years to reach a new peak at 33 countries.
• Five countries (Germany, Italy, Kazakhstan, Lithuania, and Taiwan) abandoned their nuclear programs between 1987 and 2025, one more than in mid-2024.
• Eight of the 31 nuclear countries have active reactor construction programs, two fewer than as of mid-2024.
• Twenty-three nuclear countries—three quarters—are not constructing any reactors currently (or construction is suspended); of these, two have either nuclear phaseout or no-newbuild legislation (still) in place (Spain and Switzerland). Some of these program-limitation policies, such as in Belgium, the Netherlands, and Sweden, have been revised only recently. However, while policy changes in some countries reopen or open the door for nuclear newbuild, actual work on the ground remains many years away.

1. National Nuclear Power Programs Development, 1954–2025

Sources: Various, compiled by WNISR with IAEA-PRIS, 2025

In 2024, the world nuclear fleet generated 2,677 net terawatt-hours (TWh or billion kilowatt-hours) of electricity16, (see Figure 2). After a decline in 2020, nuclear production increased by 4.2 percent in 2021, dropped by 4.4 percent in 2022, increased by 2.2 percent in 2023 and again by 2.9 percent in 2024. It is the highest production ever, the previous record being 2,663 TWh in 2006. China, with a 3.7-percent increase (compared to 11.3 percent in 2021, 2.5 percent in 2022 and 4.1 percent in 2023), produced more nuclear electricity than France for the fifth year in a row, and remains in second place—behind the U.S.—of the top nuclear power generators. Nuclear production outside China increased by 2.8 percent, to break even with the level of global production already reached in the mid-1990s prior to China’s nuclear buildout.

If global nuclear generation exceeded the previous record by modest 14 TWh, hardly more than what a single of the largest reactors can nominally generate annually, outside China, nuclear generation in 2024 was by 363 TWh below the 2006 level, a whooping 13.6-percent plunge.

Nuclear energy’s share of global commercial gross electricity generation in 2024 was almost stable at 9 percent (–0.13 percentage points)—the lowest value in four decades—and remained over 45 percent below the peak of 17.5 percent in 1996.

Nuclear’s main competitors, non-hydro renewables, grew their gross output by 14.1 percent and their share in global gross power generation increased by 1.5 percentage points to 17.3 percent.

Non-hydro renewables, including mainly solar, wind and biofuels, continued their growth, with a 12.5 percent increase, to reach a share of 8.9 percent in primary energy. While the share of non-hydro renewables is now more than two times larger than the nuclear share of 4 percent, both figures illustrate how modest the current contribution of both technologies remains in the global context.

In 2024, there were five countries—compared to nine in 2023—that increased the nuclear share in their respective electricity mix, including two “newcomer countries”, the United Arab Emirates (UAE) and Belarus, which generated nuclear power for the first time in 2020, while five decreased the share, and 22 remained at a constant level (change of less than 1 percentage point). Besides the UAE and Belarus, six countries (Argentina, China, India, Iran, South Korea, Pakistan) achieved their largest ever nuclear production. China and India started up new reactors during the year, and South Korea connected one unit to the grid in December 2023, which only impacted production in 2024.

The following noteworthy developments for the year 2024 illustrate the continued volatile operational situation of the individual national reactor fleets (see country-specific sections for details):

• Argentina’s nuclear production increased again by 16.5 percent following a 20-percent jump in the previous year and reached the level of 2021 preceding major maintenance outages in 2022.
• Belarus has seen another increase of 34 percent, after the 150-percent increase in nuclear generation in 2023, due to the startup of Belarusian-2 and improved operation of Unit 1.
• Belgium’s nuclear generation, following a 25-percent decline in 2023, partially due to the closure of one unit early in the year, dropped again by more than 5 percent, and will continue to decrease significantly with the scheduled closure of three more reactors.
• China started up three units like in each 2022 and 2021 but only one in 2023, with nuclear generation increasing 3.7 percent, the third year in a row around 4 percent, following an 11.2-percent surge in 2021.
• Finland saw a 5-percent drop following a 35 percent increase in nuclear generation in 2023 due to the beginning of the commercial operation of the largest nuclear reactor in the country, Olkiluoto-3. The country’s flagship reactor had a disappointing annual load factor of 70 percent.
• France’s nuclear generation increased again by close to 13 percent after starting recovery in 2023 (+15 percent) following the “annus horribilis” 2022 that had seen drop below 300 TWh for the first time since 1990—production however remained below 400 TWh, for the ninth year in a row.
• India’s nuclear output grew by 13.5 percent due to the startup of a new reactor early in the year.
• Iran managed with its only operating reactor to achieve an annual load factor of 80 percent for the first time and increased production by over 11 percent to a new record.
• Japan’s nuclear output increased by another 9.6 percent following the restart in the second half of 2023 of two units and further two units in 2024—14 reactors are now operational. As these latest restarts took place towards the end of the year, further production improvement is to be expected in 2025. However, nuclear represented only about 8.5 percent of total power generation in the country.
• The Netherlands’ only over 50-year-old reactor produced about 10.5 percent less power than in the previous year as the load factor dropped just below 80 percent for the first time since 2018.
• Taiwan’s nuclear output went down by more than 31.5 percent with one of two remaining reactors closing in July. Taiwan completed its nuclear phaseout and closed its last reactor in May 2025.
• The UAE connected its last reactor under construction to the grid in the first quarter and increased nuclear power generation by another 18 percent to a new record. The share of nuclear in national electricity generation reached almost 23 percent.

1. Nuclear Electricity Generation in the World... and China

Sources: WNISR with Energy Institute, 2025

Similar to previous years, in 2024, the “big five” nuclear generating countries—the U.S., China, France, Russia, and South Korea, in that order—generated 73 percent of all nuclear electricity in the world (see Figure 3, left side).

In 2002, China was 15th in terms of global production levels; in 2007, it was tenth, and it reached third place in 2016. In 2020, China became the second largest nuclear generator in the world, a position that France held since the early 1980s. That has not changed since.

1. Nuclear Electricity Generation and Share in National Power Generation

Sources: compiled by WNISR with Energy Institute, 2025

Notes: For comparison reasons, data used in this graphic are Energy Institute data17, and may differ from data used in the country sections.

Production: in TWh net; Share: calculated from gross data (nuclear and total generation), except for Armenia (IAEA-PRIS).

In 2024, the top three countries, the U.S., China, and France, remained at around 59 percent of global nuclear output, underscoring the concentration of nuclear power generation in a very small number of countries.

In many cases, even where nuclear power generation increased, the addition is not keeping pace with overall increases in electricity production, leading to a nuclear share below the respective historic maximum (see Figure 3, right side). Six of the 32 countries that operated nuclear reactors in 2024 achieved their historically largest nuclear shares in the 1980s, seven in the 1990s, six in the 2000s, two in the 2010s, eleven in the short 2020s with five for the single year of 2024.

Operation, Power Generation, Age Distribution

Since the first nuclear power reactor was connected to the Soviet power grid at Obninsk in 1954, there have been two major waves of startups. The first peaked in 1974, with 26 grid connections. The second reached a historic maximum in 1984 and 1985, just before the Chornobyl accident in 1986, reaching 33 grid connections in each year. By the end of the 1980s, the uninterrupted net increase of operating units had ceased, and in 1990 for the first time the number of reactor closures18 outweighed the number of startups.

1. Nuclear Power Reactor Grid Connections and Closures in the World

Sources: WNISR with IAEA-PRIS, 2025

Note: WNISR considers reactors closed as of the date of their last electricity production, and not as of their closure announcement (which can be made years after the reactor ceased production).

The 1995–2004 decade globally saw slightly more startups than closures (44/37), while in the decade 2005–2014, startups compensated only for two thirds of the closures (35/54) (see Figure 4).

In the decade 2015–2024, startups (69) doubled compared to the previous decade—of which over half (37) in China—but closures reached 47 none of which in China. In other words, a net decline of 15 units outside China.

Over the past two decades 2005–2024, there were 104 startups and 101 closures. Of these, 51 startups were in China which did not close any reactors. As a result, outside China, there has been a drastic net decline by 48 units over the same period (see Figure 5). As larger units were started up (totaling 97.2 GW) than closed (totaling 74 GW) the net nuclear capacity added worldwide over the 20-year period was 23.2 GW. However, since China alone added over half of the capacity (49.8 GW), the net capacity outside China declined by almost 26.6 GW.

In 2024, seven reactors were connected to the grid, three in China, and one each in France, India, the UAE and the U.S.—leaving zero active construction projects in France, the UAE, and the U.S.—and four units were closed, two in Canada and one each in Russia and Taiwan.

In the first half of 2025, one unit was connected to the grid in India—remarkably, not a single unit in China or indeed elsewhere—while two were closed, one each in Belgium and Taiwan. (See Figure 5).

1. Nuclear Power Reactor Grid Connections and Closures World/China

Sources: WNISR, with IAEA-PRIS, 2025

As of 1 July 2025, a total of 408 nuclear reactors—same number as reported for mid-2024—were operating in 31 countries—one less than in mid-2024.19 As of mid-2025, the world fleet has a total electric net operating capacity of 368.7 GW, while there were 33 units (28.6 GW) in Long-Term Outage (LTO): 19 in Japan, six in Ukraine, three in India, two each in Canada and South Korea, and one in China. As the annual statistics always reflect the status at year-end, the situation might change again by the end of 2025. As of the end of 2024, the operating capacity had reached a record 369.4 GW. The annual differences are within the one-digit GW range, to be compared with hundreds of GW range for renewable energy competitors.

The number of operating reactors remains by ten below the fleet size already reached in 1989 and by 30 below the 2002-peak (see Figure 6).

1. World Nuclear Reactor Fleet, 1954–mid-2025

Sources: WNISR with IAEA-PRIS, 2025

For many years, the capacity increased more than the number of reactors as a result of the combined effects of larger units replacing smaller ones and “uprating”. In 1989, the average size of an operational nuclear reactor was about 740 MW, in 2024 it was around 900 MW. Technical alterations raised capacity at existing plants resulting in larger electricity output, a process known as uprating.20 In the U.S. alone, the Nuclear Regulatory Commission (U.S. NRC) has approved 172 uprates since 1977. The cumulative approved uprates in the U.S. total 8 GW, the equivalent of eight large reactors. These include seven minor uprates (<2 percent of reactor capacity) approved since mid-2020, of which only one since mid-2021.21 So this is a program that is pretty much completed on most of the U.S. fleet.

A similar trend of uprates and major overhauls in view of lifetime extensions of existing reactors has been seen in Europe. The main incentive for lifetime extensions is economical, but this argument is being increasingly challenged as refurbishment costs soar and alternatives become cheaper.

The IAEA’s Operating Reactors Data Revisions

Until September 2022, the IAEA’s online Power Reactor Information System (PRIS) database counted 33 reactors as operational/operating in Japan, whereas 20 of these had not produced power since 2010–2012, and an additional three units had been shut down even since the Niigata Earthquake in 2007.

For almost a decade WNISR had been calling for an appropriate reflection in world nuclear statistics of the unique situation in Japan. The approach taken by the IAEA, the Japanese government, utilities, industry, and many research bodies as well as other governments and organizations to continue classifying the entire stranded reactor fleet in the country as “in operation” or “operational” was clearly misleading.

Faced with this dilemma, the WNISR team in 2014 decided to create a new category with a simple definition, based on empirical fact, without room for speculation: “Long-Term Outage” or LTO. Its definition:

A nuclear reactor is considered in Long-Term Outage or LTO if it has not generated any electricity in the previous calendar year and in the first half of the current calendar year. It is withdrawn from operational status retroactively from the day it has been disconnected from the grid.

When subsequently the decision is taken to close a reactor, the closure status starts with the day of the last electricity generation, and the WNISR statistics are retroactively modified accordingly.

The IAEA’s Category of “Suspended Operation”

On 16 January 2013, the IAEA moved 47 reactors in Japan, most of them shut down in the aftermath of the Fukushima events in 2011, from the category “In Operation” into “Long-term Shutdown”22 that existed in the IAEA statistical system until October 2022. Only two days later, the move was labelled a “clerical error” and the action was reversed at the request of the Japanese government.23

It is only in September 2022, that in the IAEA-PRIS database, twelve Japanese reactors24 were gradually withdrawn from the list of “Operating” or “Operational” reactors, and their status changed to “Long-term Shutdown” (LTS). By mid-October 2022, the category title was changed to “Suspended Operation” on the PRIS website25, and in November 2022, four more Japanese units26 joined the new category as well as one Indian reactor (Rajastan-1) that has not generated any power since 2004 and is considered closed by WNISR.

As of the end of 2022, the PRIS database still counted 17 Japanese reactors as “In Operation”. Whereas ten have effectively restarted since the beginning of the Fukushima disaster (also referred to as 3/11), the remaining seven have not produced any electricity since 2010–2012. Then, in April 2023, those seven units also joined the “Suspended Operation” category, followed in May 2023 by three additional Indian reactors, that have not produced power since 2018 (Madras-1) and 2020 (Tarapur-1 & -2).

The definition of the new IAEA category is as follows:

A reactor is considered in the suspended operations status, if it has been shut down for an extended period (usually more than one year) and there is the intention to re-start the unit but:

1. restart is not being aggressively pursued (there is no vigorous onsite activity to restart the unit) or

2. no firm restart date or recovery schedule has been established when unit was shutdown [shut down].

Suspended operations may be due to [due to] technical, economical, strategic or political reasons. This status does not apply to long-term maintenance outages, including unit refurbishment, if the outage schedule is consistently followed, or to long-term outages due to regulatory restrictions (licence suspension), if restart (licence recovery) term and conditions have been established. Such units are still considered “operational” (in a long-term outage). If an intention not to restart the shutdown unit has been officially announced by the owner, the unit is considered “permanently shutdown [shut down]”.27

It is important to understand that the application of this new rule modifies retroactively all of the IAEA’s statistics on operating reactors—in most cases as of day of last production—back to 2007. This dramatically modifies the IAEA’s representation of the Japanese nuclear reactor fleet’s evolution. The changes obviously also impact the IAEA’s representation of the long-term evolution of the entire global nuclear power-reactor fleet (see a detailed discussion in WNISR2023).

The differences with WNISR statistics are greatly reduced, and the remaining ones mostly relate to official closure dates, as WNISR statistics consider the end of electricity production as reference for dating closures, and not the “announcement” or “political decision” to permanently withdraw a reactor from the grid. Currently, a significant difference is that the IAEA considers the six Zaporizhzhia reactors as operational although they have not produced any power in years (see hereunder and Ukraine Focus).

IAEA vs. WNISR Assessment

WNISR’s assessment of “operating” reactors has shown significant differences with IAEA statistics since the beginning of the Fukushima disaster in 2011. However, after major changes in the PRIS statistics (see above), those differences were reduced to minor disparities during the period September 2022 to May 2023.

The following section provides a detailed explanation and justification of the differences.

Figure 7 presents the evolution of the number and capacity of the world reactor fleet “in operation” as reported by the IAEA vs. WNISR.

As of 1 July 2025, the evolution of the world nuclear fleet in the IAEA-PRIS statistics shows a peak of 440 reactors operating in 2005, whereas, as of the end of 2024, the operating capacity reached 377 GW, a new record, marginally over the previous peak of 374 GW reached in 2018, according to the new IAEA data.

1. World Nuclear Reactor Fleet – IAEA vs WNISR, 1954–July 2025

Sources: IAEA-PRIS and WNISR, 2025

Notes: The IAEA data used for this graph includes at least three reactors that have been later withdrawn from the PRIS statistics for operating reactors (Niederaichbach, VAK-Kahl and HDR Großwelzheim, in Germany, now only appearing as “Decommissioning Completed”). On the other hand, the Swiss research reactor in Lucens is not included. Reactors classified as in “Suspended Operation” by the IAEA are not represented here.
Although the total number of reactors in operation according to WNISR statistics has always remained, albeit slightly, inferior to IAEA-PRIS data, it contains several Chinese reactors not accounted for in PRIS (see below).

In the WNISR statistics, which consider reactors closed from the day they stop producing electricity and systematically apply the LTO status to reactors not operating for a certain period, a maximum number of 438 operating reactors was reached as soon as 2002, and again in 2005. At the end of 2024, with a balance of plus 5.3 GW between closed and newly started-up reactors during the year plus a balance of three fewer reactors in LTO, the operating capacity reached a new maximum of 369.4 GW.

Although not the only case, the Japanese fleet still provides the main and more visible differences between the two datasets, especially over the decade 2011–2020. This applies both to reactors that did not produce electricity for many years before they returned to service (designated as “LTO later restarted” or “Restarted from LTO”), or which were declared permanently closed years after they stopped producing electricity (“Closed at a later date”).

Applying this definition to the world nuclear reactor fleet, as of 1 July 2025, leads to classifying ten units considered “in operation” by the IAEA as LTO:

• Bruce-3 and Darlington-4 in Canada under refurbishment since March and July 2023 respectively (see section on Canada).
• Kori-2 and Kori-3 in South Korea, shut down in April 2023 and September 2024 respectively, after 40 years of operation and the expiration of their operating license. Currently, they are expected to be restarted at an unknown date, pending inspection results and potential upgrading; therefore they are considered in LTO (see South Korea Focus).28
• The six reactors at Zaporizhzhia in Ukraine that did not produce any electricity in 2023 and 2024 and are considered in LTO as of the end of 2022.

As of mid-2025, the IAEA classified 19 reactors in Japan and four in India as suspended. With the exception of one reactor in India (Rajasthan-1) considered closed since the end of 2004, WNISR classifies all of them in LTO.

On the other hand, WNISR statistics do include additional reactors in China:

• Shidao-Bay-1: The IAEA considers the two 100-MW modules as one reactor as they drive a single 200-MW turbine. WNISR considers that each module is a separate reactor.
• Shidaowan Guohe One-1: This is the first of two CAP1400 at Shidao Bay that started operating in 2024, a second unit is still under construction. For unknown reasons, the two units are not listed in the IAEA database.
• China Experimental Fast Reactor (CEFR): The IAEA has simply deleted the file for the reactor without any indication of reasons. Chinese sources have argued it should have never been in the IAEA’s PRIS database in the first place as it is to be considered an experimental reactor. However, as this is a grid-connected nuclear power reactor, it is considered as such by WNISR. Its current operational status is uncertain. In the absence of operational data, WNISR considers it in LTO as of May 2023 (but still operating as of December 2022).29

The biggest difference between IAEA-PRIS and WNISR is found as of the end of 2012, with 29 units less operating according to WNISR criteria: the IAEA-PRIS counts 30 reactors (detailed in Table 1) that are not considered operating according to WNISR, but on the other hand has retrieved the Chinese CEFR it previously considered operational at that date.

1. WNISR Rationale for the Classification of 30 Reactors as Non-Operational as of end 2012

Countries	Officially Closed at a Later Date 21 Reactors		Restarted from LTO 9 Reactors
	Reactors that last produced electricity in (or prior to) 2012, officially closed after 2012 (either considered closed by WNISR as early as 2012, or after a certain period in LTO). Most of those reactors were considered “in operation” for many years before their official closure date.		Reactors in LTO as of December 2012 Restarted prior to 1 July 2023
	Reactors considered closed in 2012	Reactors in LTO prior to closure
Japan	6 Reactors Fukushima Daiichi 5–6 Fukushima Daini 1–4 Officially Closed in 2013 and 2019	11 Reactors Last production in 2010–2012 Officially closed 2015–2019	8 Reactors Restarted 2015–2021
South Korea			1 Reactor Wolsong-1, Restarted in 2015
Spain	1 Reactor Santa Maria de Garoña Last production in 2012 Officially Closed in 2017*
U.S.	3 Reactors San Onofre-2 & -3 Last production in 2012 Officially closed in 2013 Crystal River-3 Last production in 2009 Officially closed in 2013

Sources: IAEA-PRIS and WNISR, 2024

Note: *Garoña was subsequently considered in “Suspended Operation” during 2013–2016 by the IAEA until its official closure.

The differences between the IAEA and WNISR are not limited to the effects of the Fukushima disaster. Even prior to 3/11, WNISR and IAEA-PRIS data had differences, reaching up to 10 units at the end of some years. These differences were mainly due to the definition of the closure date that the IAEA either sets at last production or as closure-decision date while WNISR systematically applies the day of last electricity generation (when available). Other reasons for differences lie in the IAEA’s delays to classify reactors in “suspended operation” and that WNISR considers reactors undergoing refurbishment in LTO as soon as they meet the LTO criteria, or as soon as they are shut down (with expected restart) in case their license has expired (as it is the case in South Korea for example).

Overview of Current Newbuild

As of 1 July 2025, 63 reactors were considered as under construction30—including 32 units in China—four more than the WNISR reported a year ago and six fewer than in 2013 (see Figure 8 and Figure 9). Of the 69 reactors listed at the end of 2013, six projects were subsequently abandoned or suspended.

Nine in ten of all reactors are being built in Asia or Eastern Europe (see Building vs. Vendor Countries). Eleven countries are building nuclear plants, down from 16 in as of mid-2023 and 13 as of mid-2024. While Pakistan entered the category with the launch of a new construction at one of its existing sites, three countries dropped off the list: France, where the latest and last reactor under construction was connected to the grid; Argentina, where the CAREM reactor construction has been abandoned, and Japan where no active construction is underway at both construction sites that had been previously listed at some point in time.

Only four countries—China, India, Russia and South Korea—have construction ongoing at more than one site, and three countries only have a single reactor under construction (see Table 2 and Annex 5 for details). Between mid-2024 and mid-2025, construction of ten units was launched worldwide, seven in China and one each in Pakistan, Russia, and South Korea.

The 63 reactors listed as under construction represent just over one quarter of the 234 units—totaling more than 200 GW—listed in 1979. However, 48 of those projects listed at the time were never finished (see Figure 8). The year 2005, with 26 units listed as under construction, was the lowest since the early nuclear age in the 1950s.

1. Nuclear Reactors “Under Construction” in the World

Sources: WNISR, with IAEA-PRIS, 2025

Notes: This figure includes construction of two CAP1400 reactors at Rongcheng/Shidaowan, although their construction has not been communicated by the IAEA, one of which has been connected to the grid in October 2024 (see China Focus).

Compared to mid-2024, the number of units under construction in the world increased from 5831 to 63  in mid-2025 with the combined capacity increased by 6.7 GW net to 65.2 GW, with an average size of 1 GW.

1. Nuclear Reactors “Under Construction” – China and the World

Sources: WNISR with IAEA-PRIS, 2025

Building vs. Vendor Countries

As of mid-2025, China hosts 32 or more than half of all reactors under construction in the world (63). So far, Chinese companies have only exported to Pakistan, where they started building another reactor in December 2024.

It is remarkable that, as of mid-2025, there is not a single active construction on the entire American continent, from Alaska to Cape Horn, including the United States, that hosts the world’s largest nuclear power fleet (see Figure 10).

Russia is in fact largely dominating the international market as a technology supplier with 27 units under construction in the world, as of mid-2025, of which seven at home and 20 in seven different countries, including four each in China, Egypt, India, and Türkiye, and two in Bangladesh. It remains uncertain to what extent these projects are impacted by the various layers of sanctions imposed on Russia following its invasion of Ukraine.

Besides Russia’s Rosatom, there is only France’s EDF (Électricité de France) and China’s CNNC (China National Nuclear Corporation) presently building abroad (see Table 2).

1. Nuclear Reactors “Under Construction” by Technology-Supplier Country

Sources: WNISR with IAEA-PRIS, 2025

1. Nuclear Reactors “Under Construction” (as of 1 July 2025)32

Country	Units (Domestic Design)	Other Vendor	Capacity (MW net)	Construction Start	Grid Connection	Units Behind Schedule
China	32 (28)	Russia: 4	34 512	2017 – 2025	2025 – 2030	1 (2?)
Russia	7 (7)	–	5 110	2018 – 2025	2025 – 2032	3 (5?)
India	6 (2)	Russia: 4	4 768	2004 – 2021	2026 – 2028	6
Türkiye	4 (–)	Russia: 4	4 456	2018 – 2022	2026 – 2029	4
Egypt	4 (–)	Russia: 4	4 400	2022 – 2024	2028 – 2031	-
South Korea	3 (3)	–	4 020	2017 – 2025	2026 – 2032	2
Bangladesh	2 (–)	Russia: 2	2 160	2017 – 2018	2025 – 2027	2
U.K.	2 (–)	France: 2	3 260	2018 – 2019	2030 – 2031	2
Slovakia	1 (–)	Russia: 1(a)	440	1985	2025	1
Iran	1 (–)	Russia: 1	974	1976	2029	1
Pakistan	1 (–)	China: 1	1 117	2024	2030
Total	63		65 217	1976 – 2025	2025 – 2032	22 (26?)
Total per Vendor Country: Russia: 27 – China: 29 – South Korea: 3 – France: 2 – India: 2

Sources: Various, compiled by WNISR, 2025

(a) - Mochovce -4 is a Russian VVER design being completed by a Czech-led consortium.

This table does not contain suspended or abandoned constructions. It does include construction of one CAP1400 reactor at Rongcheng/Shidaowan, although that has not been communicated by the IAEA (see China Focus) as well as two floating reactors of Russian design to be deployed in Russia—thus counted under Country-Russia, but with the barges built in China, which are likely also delayed.

Construction Times

Construction Times of Reactors Currently Under Construction

A closer look at projects listed as “under construction” as of 1 July 2025 illustrates the level of uncertainty and problems associated with many of these projects, especially given that most builders still claim a five-year construction period in their project proposals:

• For the 63 reactors being built, an average of 5.3 years has passed since construction start—slightly lower than the mid-2024 average of 5.9 years of the fleet under construction then—and many remain far from completion.
• All reactors under construction in at least six of the 11 building countries have experienced often year-long delays. Over one third (23) of the building projects are documented to be delayed. Most of the units—26 of 40—which are nominally being built on-time (yet) were begun within the past three years (July 2022–July 2025) and 30 of the 40 are in China, making it difficult to assess whether they are on schedule. Significant uncertainty remains over construction in China because of lack of systematic access to information.
• It remains also unclear what will happen with Russian designed and/or implemented projects in six other countries, as sanctions have or will likely have an impact on supply chains. Russian projects in five of the seven countries are already documented to be delayed (Bangladesh, India, Iran, Slovakia, Türkiye), some of them due to sanctions’ impacts.
• Of the 22–26 constructions behind schedule, at least fourteen—in Bangladesh, India, Iran, Russia, Slovakia, South Korea and Türkiye—have reported increased delays, three have confirmed delay indications, and one delay has been identified for the first time over the past year.
• WNISR2023 noted a total of 14 reactors scheduled for startup in 2024. At the beginning of 2024, 14 were still planned to be connected to the grid (including three pushed back from 2023 to 2024), but only half made it to generate first power, while the commissioning of the other seven was delayed at least into 2025.
• The initial construction start of the Mochovce-4 reactor in Slovakia dates back 40 years and its grid connection has—again—been further delayed, currently to later in 2025. Bushehr-2 in Iran originally started construction in 1976, almost 50 years ago, and resumed construction in 2019 after a 40-year-long suspension. Grid connection is currently scheduled for 2029.
• Two additional reactors have been listed as “under construction” for a decade or more: the Prototype Fast Breeder Reactor (PFBR) and Rajasthan-8 in India.

The actual lead time for nuclear plant projects includes not only the construction itself but, in most countries, also lengthy political and legal processes, licensing procedures, complex financing negotiations, site preparation, and other infrastructure development.

Construction Times of Past and Currently Operating Reactors

Since the beginning of the nuclear power age, there has been a clear global trend towards increasing construction times. National building programs were faster in the early years of nuclear power, when units were smaller, and safety and environmental regulations were less stringent.

As Figure 11 illustrates, average times between construction start and grid connection of reactors completed in the 1970s and 1980s were quite homogeneous, while in the past two decades they have varied widely.

1. Average Annual Construction Times in the World

Sources: WNISR with IAEA-PRIS, 2025

In the three years 2022–2024, there were 19 startups in 10 countries after an average 10.8-year construction time. None of them was on-time, delays varied from a few weeks to many years (see Figure 12). The longest construction time was seen for Mochovce-3 at 38 years in total or at 22 years net, taking into account the 16-year-long construction suspension, followed by the French Flamanville-3 European Pressurized Water Reactor (EPR) at 17.1 years, and the Olkiluoto-3 EPR at 16.6 years. The longest construction time experienced in China, overall a significantly better performer than its foreign competitors, was seen for the second of the High-Temperature Reactor Pebble-bed Modules (HTR-PM) at Shidao Bay at 9–10 years. The seven units completed in 2022–2024 in China took on average 6.8 years to build.

The seven units that started up in the world in 2024 were completed after an average construction time of 9.6 years. The mean time from construction start to grid connection for the five reactors started up in 2023 was 14.9 years, almost six years more on average than the nine years for the units started up in 2022. This was mainly due to Mochovce-3 in Slovakia.

Only one unit began power generation in the first half of 2025, Rajasthan-7 in India, after 13.7 years of construction.

1. Delays for Units Started Up, 2022–2024

Sources: Various, compiled by WNISR, 2025

Notes: Expected construction time is based on grid connection data provided at construction start when available; alternatively, best estimates are used, based on commercial operation, completion, or commissioning targets.

At Shidao Bay, the HTR plant, where construction started in 2012, has two reactor modules onsite and is therefore counted as two units as of WNISR2020. Grid connection of the first unit of the twin reactors officially took place on 20 December 2021. No date was provided for startup of the second reactor, which WNISR considers operating by the of end-2022 with total construction time set at 10 years.

The longer-term perspective confirms that short construction times remain the exceptions. Eleven countries completed 69 reactors over the decade 2015–2024—of which 37 in China alone—with an average construction time of 9.4 years (see Table 3), less than the 9.9 years of mean construction time in the decade 2014–2023, but equal to the decade 2013–2022. The construction durations from the beginning of concreting of the foundations of the reactor building to first grid connection have had better and worse years but over a decade they have averaged out around 10 years for over a decade, with a broad range between countries and between projects inside individual countries.

1. Duration from Construction Start to Grid Connection, 2015–2024

Construction Times of 69 Units Started Up 2015–2024
Country	Units	Construction Time (in Years)
		Mean Time	Minimum	Maximum
China	37	6.3	4.1	10
Russia	8	16.2	8.1	35.1
South Korea	5	8.7	6.4	10.5
Pakistan	4	5.6	5.5	5.8
UAE	4	8.3	8.0	8.6
India	3	12.5	10.1	14.2
U.S.	3	21.5	10.1	42.8
Belarus	2	8.0	7.0	8.9
Finland	1	16.6	16.6
France	1	17.1	17.1
Slovakia	1	38.1	38.1
World	69	9.4	4.1	42.8

Sources: Various, compiled by WNISR, 2025

Construction Starts and Cancellations

The number of annual construction starts in the world peaked in 1976 at 44, of which 11 projects were later abandoned. In 2010, there were 15 construction starts—including 10 in China—the highest level since 1985 (see Figure 13 and Figure 14). That number dropped to five in 2020 (including four in China), while building started on ten units in 2021 (including six in China), as well as in 2022 (including five in China). Six constructions started in 2023, of which five in China and one in Egypt implemented by the Russian nuclear industry. Nine constructions started in 2024, of which six in China, one in Russia, one in Pakistan implemented by Chinese companies, and one in Egypt implemented mainly by Russian companies. In other words, over the five years between January 2020 and the end of 2024, all of the 40 construction starts in the world were either implemented by the Chinese or the Russian nuclear industries.

Five reactors got underway in the world in the first half of 2025, three of them in China, one in Russia, and one in South Korea. So, Chinese and Russian government-owned or -controlled companies launched all 44 of the 45 reactor constructions in the world over the 66-month period from the beginning of 2020 to mid-2025.

1. Construction Starts in the World

Sources: WNISR with IAEA-PRIS, 2025

Over the decade 2015–2024, construction began on 67 reactors in the world, of which way over half (39) in China. As of mid-2025, 13 of those units had started up, while 54 remain under construction.

Seriously affected by the Fukushima events, China did not start any construction in 2011 and 2014 and began work on “only” seven units in total in 2012 and 2013. While Chinese utilities started building six more units in 2015, the number shrank to two in 2016, only a demonstration fast reactor in 2017, none in 2018, but four each in 2019 and 2020, six in 2021, five each in 2022 and 2023, six in 2024 and three in the first half of 2025 (see Figure 14). While this increase represents a sign of the restart of large-scale reactor building in China, the level continues to remain far below expectations. The five-year plan 2016–2020 had fixed a target of 58 GW operating and 30 GW under construction by 2020. As of the end of 2020, China had 49 units with 51 GW (47.5 GW net) operating, one reactor in LTO (CEFR), and 17 units (17.5 GW gross / 16 GW net) under construction, much lower than the original target.

With an operating capacity of around 61 GW as of mid-2025, the 14th Five-Year Plan (2021–2025) target of 70 GW operational capacity is also out of reach. At the end of 2024, 59 reactors with a total capacity of 56.7 GW were operating and 29 units (31 GW) were under construction. According to plans, 4.5 GW are scheduled to come online in 2025, but no new reactor started up in the first half of the year (for details and references, see China Focus).

1. Construction Starts in the World/China

Sources: WNISR with IAEA-PRIS, 2025

Experience shows that having an order for a reactor or even having a nuclear plant at an advanced stage of construction, is no guarantee of ultimate grid connection and power production. The two V.C. Summer units, abandoned in July 2017 after four years of construction and following multi-billion-dollar investment, are only one example in a long list of failed significantly advanced nuclear power plant projects. Another recent example is the Small Modular Reactor (SMR) CAREM-25, an Argentinian domestic design, that was abandoned in 2024 after ten years of construction.

French Alternative Energies and Atomic Energy Commission (CEA) statistics through 2002 indicate 253 “cancelled orders” in 31 countries, many of them at an advanced construction stage (see also Figure 15). The United States alone accounted for 138 of these order cancellations.33 Unfortunately, the data series does not distinguish between “orders” and “constructions” and it has been discontinued.

Of the 817 reactor constructions launched since 1951, at least 95 units in 20 countries had been abandoned or suspended, as of 1 July 2025. This means that 11.5 percent—or one in nine—of nuclear constructions have been abandoned.

Close to three-quarters (66 units) of all cancelled projects were in four countries alone—the U.S. (42), Russia (12), Germany and Ukraine (six each). Some units were 100-percent completed—including Kalkar in Germany and Zwentendorf in Austria—before it was decided not to operate them.

1. Cancelled or Suspended Reactor Constructions

Sources: Various, compiled by WNISR, 2025

Notes: The shaded colors indicate the first suspension in the case of reactors where construction was suspended several times (Brazil and Japan).

This figure only includes constructions that had officially started with the concreting of the base slab of the reactor building. Many more projects have been cancelled at earlier stages of construction/site preparation and equipment manufacturing.

Operating Age

In the absence of significant, successful newbuild over many years, the average age (from grid connection) of operating nuclear power plants has been increasing for 40 years, since 1984, and as of mid-2025 stands at 32.4 years, up from 32 years in mid-2024 (see Figure 16).34

A total of 266 reactors, two-thirds of the world’s operating fleet, have operated for 31 or more years, including 141—more than one in three—for at least 41 years.

1. Age Distribution of Operating Reactors in the World

Sources: WNISR, with IAEA-PRIS, 2025

In 1990, the average age of the operating reactors in the world was 11.3 years; in 2000, it was 18.8 years, and it stood at 26.3 years in 2010. The leading nuclear nation also has the oldest reactor fleet of the top-five nuclear generators. The average age of reactors in the U.S. passed 40-years in 2020 and reached 43.2 years as of the end of 2024. France’s fleet reached almost 39 years. Russia’s fleet age peaked in 2017 and declined for a few years before increasing again starting in 2020, and its average fleet age of 31 years, as of the end of 2024, caught up with that of 2017. South Korea’s reactors reached 23.7 years (including two reactors in LTO); the operating fleet, with an average age of 22.3 years remained almost half as old as the U.S. fleet. And China, the outrunner, had an average fleet age of only 10.6 years. (See Figure 17).

1. Reactor-Fleet Age of Top 5 Nuclear Generators

Sources: WNISR with IAEA-PRIS, 2025

Many nuclear utilities envisage average reactor lifetimes of beyond 40 years up to 60 and even 80 years. In the U.S., reactors are initially licensed to operate for 40 years, but nuclear operators can request a license renewal from the Nuclear Regulatory Commission (NRC) for an additional 20 years. An initiative to allow for 40-year license extensions in one step was terminated in June 2021 after NRC staff recommended that the Commission “discontinue the activity to consider regulatory and other changes to enable license renewal for 40 years.”35

As of mid-2025, 86 of the 94 operating U.S. units had received a first 20-year license extension, applications for four further reactors were under NRC review, including the Initial License Renewal application for the Diablo Canyon units, that was slated for closure and is now planned to be restarted.36

As of 1 July 2025, the NRC had granted Subsequent Renewed Operating Licenses to thirteen reactors, which permit operation from 60 to 80 years. A further twelve reactors have their applications still under review, and owners have notified of their intentions to submit applications for a further 29 reactors between 2025 and 2034, including three currently retired reactors See Extended Reactor Licenses in United States Focus for details and references.

Only nine of the 41 units that have been closed in the U.S. had reached 40 years on the grid. All nine had obtained licenses to operate up to 60 years but were closed long before, mainly for economic reasons. In other words, almost one quarter of the 135 reactors connected to the grid in the U.S. never reached their initial design lifetime of 40 years. Only one of those already closed had just reached 50 years of operation (Palisades, closed after 50.4 years, and currently undergoing a potential restart process). The mean age at closure of those 41 units was 22.8 years.

On the other hand, of the 94 currently operating plants, 57 units have already operated for 41 years or more, of which 20 have been on the grid for 51 years or more; thus, two thirds of the units with license renewals have entered the lifetime extension period, and that share is growing rapidly with the mid-2025 mean age of the U.S. operational fleet exceeding 43.7 years (see Figure 46).

Many countries have no specific time limits on operating licenses. In France, for example, reactors must undergo in-depth inspection and testing every decade against reinforced safety requirements. The French reactors have operated for over 39 years on average, as of mid-2025.

EDF’s approach to lifetime extension has been reviewed by the Nuclear Safety Authority (ASN) and its Technical Support Organization. In February 2021, ASN granted a conditional generic agreement to lifetime extensions of the 32 reactors of the 900 MW series. Lifetime extensions beyond 40 years require reactor-specific licensing procedures involving public inquiries in France and transborder consultations. For an assessment of the status of fourth decennial inspections see Decennial Inspections – Lifetime Extensions in France Focus.

Recently commissioned reactors and units under construction, e.g., in South Korea and the U.K. do or will seek a 60-year operating license from the start.

1. Age of World Nuclear Fleets

Sources: WNISR with IAEA-PRIS, 2025

Note: This figure only takes into account reactors operating as of 1 July 2025, thus excluding reactors in LTO, in particular Tarapur-1 & -2 in India, that have passed 50 years.

Figure 18 shows that the average fleet age in 22 of the 31 countries that operate nuclear reactors as of mid-2025 is over 30 years, and in ten countries over 40. Two in three, that is 22 of the countries have been operating one or more reactors for more than 40 years, but only seven countries operate reactors that are over 50 years.

In assessing the likelihood of reactor fleets being able to operate for 50 or 60 years, it is useful to compare the age distribution of reactors that are currently operating with the 218 units that have already closed (see Figure 16 and Figure 19). In total, 103 of these units operated for 31 years or more, of which 45 reactors operated for 41 years or more. Many units of the first-generation designs only operated for a few years. The mean age of the closed units is about 29 years.

At the same time, for the first time, in 2024 reactors closed that have been operating for 51 years and more (Pickering-1 and -2 in Canada). Only few units have been closed exceeding 50 years of operational lifetime, e.g. Doel-1 in Belgium at 50.5 years and Palisades in the U.S.

1. Age Distribution of Closed Nuclear Power Reactors

Sources: WNISR with IAEA-PRIS, 2025

While the operating time prior to closure has clearly increased continuously, the mean age at closure of the 28 units taken off the grids in the five-year period between 2020 and 2024 was just 43.2 years (see Figure 20).

As a result of the Fukushima nuclear disaster (also referred to as 3/11), many analysts have questioned the wisdom of operating older reactors. The Fukushima Daiichi units (1 to 4) were connected to the grid between 1971 and 1974. The license for Unit 1 had been extended for another 10 years in February 2011, just one month before the catastrophe started unfolding. Four days after the initial events in Japan, the German government ordered the closure of eight reactors that had started up before 1981, two of which were already closed at the time and never restarted. The sole selection criterion was operational age—30 years or more. Other countries did not adopt the same approach, but clearly the 3/11 events in Japan had an impact—at least temporarily—on previously assumed extended lifetimes in other countries. Some of the main nuclear countries closed their oldest units, at the time, before or long before age 50, including Germany at age 37, South Korea at 40, Sweden at 46. France closed its two oldest units in Spring 2020 at age 43. The U.S. closed its oldest unit, Palisades, at age 50 in 2022 and is now considering reopening it.

1. Nuclear Reactor Closure Age

Sources: WNISR with IAEA-PRIS, 2025

Lifetime Projections

Nuclear operators in many countries continue to implement or prepare for lifetime extensions. In previous years, WNISR had modeled two lifetime projections. A first scenario assuming a general lifetime of 40 years for worldwide operating reactors as that, explicitly or implicitly, corresponds to the design lifetimes of most operating reactors; and a second scenario, the Plant Life Extension or PLEX Projection (see Figure 21), which takes into account all already-authorized lifetime extensions as of mid-year and assumes that the respective reactors will operate until the expiration of their license—a conservative assumption considering empirical evidence from the past.

As operational lifetime extensions or license renewals beyond 40 years have increasingly become the norm, as of this edition WNISR drops the 40-year scenario.

Generalized lifetime extensions—far beyond 40 years—are clearly the objective of the international nuclear power industry, and, especially in the U.S., there are numerous, increasingly successful attempts to obtain subsidies for uneconomic nuclear plants in order to keep them on the grid, or even restart units that have been closed (see Subsidies and Financing for Nuclear Power in United States Focus). Expensive operational lifetime extensions are now also planned in Europe, e.g., for two reactors for ten years in Belgium.

The extension of the operation of the current nuclear fleet is essential to maintaining at least the status quo as not enough reactors are coming on line to compensate for closures. Developments in Asia, including in China, do not fundamentally change the global picture. Reported ambitions for China’s targets for installed nuclear capacity have fluctuated in the past. While construction starts have picked up speed again since 2021, Chinese medium-term ambitions appear significantly lower than anticipated in the pre-3/11 era.37

The lifetime projection allows for an evaluation of the number of reactors and respective power generating capacity that would have to come online over the next decades to offset closures and simply maintain the same number of operating plants and level of capacity, if all units were closed after their present licensed lifetime extension.

1. The PLEX Projection (not including LTOs)

Sources: Various sources, compiled by WNISR, 2025

Notes

Restarts or closures amongst the 33 reactors in LTO as of 1 July 2025 are not represented; however, at least some are expected to be restarted (and later closed, after 2050 in some cases).

In the case of reactors that have reached 40 years of operation prior to 2025, the projection uses the end of their licensed lifetime, including 80 years for 13 reactors in the U.S, where the Subsequent License Renewal Applications have been approved for a further 20 years of operation.

In the case of the French 900 MW-fleet the projection uses the scheduled date of their 4th periodic safety review. In case this deadline is or will be passed by the end of 2025, we apply a “de facto” 10-year extension (operation to 50 years) or the deadline for the fifth review when available.

In the PLEX Projection, all currently licensed lifetime extensions and license renewals are implemented in full, and all construction sites are completed. For all other units, a 40-year lifetime projection is maintained, unless a firm later closure date has been authorized. By the end of 2025, the net number of operating reactors would slightly decline, and the operating capacity would marginally increase (-4 units /+0.8 GW).

In the remaining years to 2030, the net balance would be just in the positive for the year 2026, then drop into negative terrain in the years 2027–2029 with a sharp decrease in 2030; overall, over the period 2025–2030, 44 new reactors or 26 GW in addition to the 59 units already under construction and scheduled to start up prior to 2030 would have to be built and commissioned (or units in LTO restarted) to replace expected closures.

Taking into account the already licensed lifetime extensions, the PLEX-Projection would mean for the remaining years to 2030, the need to ramp up the annual startup rate of the past decade by a factor of 2.5 from 6.9 to 17.3 units—including the reactors already expected to start up by 2030 (see Figure 21)—only to maintain the status quo. However, probably at least half of the 104 reactors now projected to close between 2025 and 2030 are seeking and likely to secure a lifetime extension beyond 2030.

As documented in detail above, construction starts have not been picking up over the past decade. Between the beginning of 2020 and mid-2025, a total of 45 constructions were launched around the world, of which 30 implemented by Chinese companies (including one in Pakistan) and 14 by the Russian industry in various countries, plus one in South Korea, thus an average of 8.2 units per year were launched. Based on empirical evidence, it is unlikely if not impossible that any substantial number of reactors will come online by 2030 that are not under construction as of mid-2025. In other words, newbuild will not be sufficient, only further operational lifetime extensions will allow for the world nuclear fleet not to decline significantly by 2030 and thereafter.

Focus Countries

China Focus

Overview

As of mid-2025, China operated 59 nuclear reactors with a combined capacity of 56.7 GW. This count exceeds the IAEA’s tally of 57 because the WNISR records the Shidao Bay project’s twin High-Temperature Reactor Pebble-bed Modules (HTR-PM) as two separate reactors and includes the CAP1400 Shidaowan Guohe One-1.38 (see Table 23 and Table 24 in Annex 2).

In 2024, nuclear plants generated a record 418.4 TWh net electricity, a 2.9 percent increase from 2023.39 Two 1.2-GW Hualong One reactors started up over the year, Fangchenggang-4 in Guangxi Province and Zhangzhou-1 in Fujian Province, in addition to the CAP1400 at Shidao Bay.

According to the China Nuclear Energy Association, despite higher output, nuclear’s share of China’s total electricity production slightly slipped from 4.9 percent in 2023 to 4.7 percent in 2024, (Energy Institute data indicate a 3.7-percent increase in net production and a drop from 4.7 percent to 4.5 percent of the nuclear share).40 The remarkable share decline occurred because China’s electricity consumption grew by 6.8 percent or 627 TWh—significantly larger than Germany’s total annual demand—to a total of over 9,850 TWh, and the country added a combined 357 GW of solar and wind capacity (278 GW and 79 GW, respectively) in the same year compared to just 3.5 GW of new nuclear.41

Some reactors significantly underperformed in 2024:

• The twin HTR-PM reactors were operational for only about one fifth of the time, partially due to ongoing testing.42 It is unclear what other factors have contributed to the low performance.
• Fuqing-4 was offline for more than one third of the year due to repairs after an incident involving metal debris from a scratched spent fuel assembly support structure in the cooling pool.43
• Daya Bay-1, a Framatome designed 900-MW reactor, was the first reactor in China subject to a major 200-day overhaul after three decades of operation. It was originally connected to the grid in August 1993. The comprehensive upgrading involved 200 upgrades including the digitalization of the instrumentation and control system.44
• Taishan-2, an EPR, had a load factor of only 69.1 percent, but no official explanation was provided for this. The twin unit Taishan-1 had a better year by comparison—offline between August 2021 and August 2022 and again for almost all of 2023 mainly for fuel assembly repairs—rebounded in 2024 and likely accounts for the lion’s share of the overall production increase.45 The two 1660-MW EPRs are the largest units in the Chinese fleet.

Construction began on six new reactors in 2024:

• Lianjiang-2 in Guangdong Province (CAP1000, 1224 MW),
• Ningde-5 in Fujian Province (Hualong One, 1200 MW),
• Shidaowan-1 in Shandong Province (Hualong One, 1134 MW),
• Xudapu-2 in Liaoning Province (CAP1000, 1000 MW), and
• Zhangzhou-3 and Zhangzhou-4 in Fujian Province (both Hualong One, 1129 MW each).

Another three construction starts took place in the first half of 2025:

• Lufeng-1 in Guangdong Province (CAP1000, 1000 MW),
• Shidaowan-2 in Shandong Province (Hualong One, 1134 MW), and
• Taipingling-3 in Guangdong Province (Hualong One, 1209 MW).

China also advanced with its only active international nuclear project. On 30 December 2024, it poured first concrete for an 1100-MW Hualong One reactor, Unit 6 at Pakistan’s Chashma nuclear power complex46—the third unit of the 1200-MW class China has exported to its neighbor. Previously, between 1993 and 2017, China had delivered four units of the 300-MW class to Pakistan, which remains the sole country hosting Chinese-built reactors.

Domestically, construction approvals have ramped up and stabilized with six reactors approved in 2019, four in 2020, five in 2021, and ten each in 2022 and 2023.47

In 2024–2025, the Chinese State Council approved two major batches of new projects. On 19 August 2024, it greenlit five projects totaling 16 reactors—including six SMR modules—at a reported cost of CNY2024220 billion (US$202430.6 billion):48

• Xuwei Phase I in Jiangsu Province (two Hualong Ones and a 660-MW high-temperature gas-cooled reactor plant with six modules),49
• Lufeng Phase I in Guangdong Province (two CAP1400s),50
• Zhaoyuan Phase I in Shandong Province (two Hualong Ones),51
• San’ao Phase II in Zhejiang Province (two Hualong Ones), and
• Bailong Phase I in Guangxi Province (two CAP1000s).52

On 27 April 2025, a second batch of approvals included five projects (10 reactors) costing a reported estimated total of over CNY200 billion (US$27.8 billion):53

• Fangchenggang Phase III in Guangxi Province (two Hualong Ones),
• Taishan Phase II in Guangdong Province (two Hualong Ones),
• Sanmen Phase III in Zhejiang Province (two Hualong Ones),
• Haiyang Phase II in Shandong Province (two CAP1000s), and
• Xiapu PWR Phase I in Fujian Province (two Hualong Ones, co-located with two 600-MW sodium-cooled fast reactors with uncertain status, see hereunder).

Sanmen Phase I and II in Zhejiang Province used the AP-1000 design or its Chinese adaptation—the CAP1000—in the case of the two latest units, while Taishan Phase 1 in Guangdong Province employed the EPR; both sites will now switch to Hualong One for future units.

In the decade 2015–2024, China expanded the number of operating reactors from 31 to 59 while more than doubling the operating capacity of its fleet from 26.8 GW to 56.7 GW (see Figure 22).

1.  The Expansion of the Chinese Nuclear Fleet, 1991–2025

Sources: WNISR with IAEA-PRIS and NNSA, 2025

With the startup of four AP-1000 reactors in 2018, China began operating its first so-called Generation III reactors.54 The first Hualong One, China’s domestic Generation III (Gen III) design, was connected to the grid in 2020, followed by one each in 2022 and 2023 and two in 2024.

China’s Nuclear Sector and Ecosystem

Since the launch of major economic reforms in 1978, China has gradually expanded its civilian nuclear use, actively pursuing international partnerships to acquire advanced technology.55

In 1985, construction began on China’s domestically designed 300-MW class Qinshan-I reactor. In 1986, under Deng Xiaoping, China purchased two French 900-MW reactors for the Daya Bay project, marking its first major step into commercial civilian nuclear power. A year later, construction began on the Daya Bay site,56 in spite of the largest opposition movement to a nuclear power project in history to that date collecting over one million signatures within a few months.57

While China possesses abundant coal, it has limited oil and gas reserves. Despite heavy reliance on coal for electricity generation, the country became a net importer of oil in 1993 and of natural gas in 2009.58 In the early 2000s, as economic growth surged, China faced worsening power shortages, logistical bottlenecks, and environmental degradation. Behind the scenes, nuclear energy has been touted as a cleaner, more stable alternative. Under Premier Wen Jiabao (2003–2013), the country began accelerating the expansion of its nuclear fleet.

At the same time, China remained committed to achieving technological self-sufficiency. To facilitate quicker knowledge transfer, it widely opened its nuclear market to foreign companies in the mid-1990s. Simultaneously, Chinese vendors began developing their own domestic Generation III PWRs, building on imported designs from France and the United States. In the early 2010s, Beijing merged CNNC’s and CGN’s competing reactor designs into a unified national model: the Hualong One.

China has not limited itself to PWRs. It has actively pursued alternative technologies, including high-temperature gas-cooled pebble-bed reactors, sodium-cooled fast reactors, and thorium-based molten salt reactors. These efforts often revive earlier Western designs abandoned elsewhere but seen in China as promising for its long-term energy strategy.

State-Owned Giants: China National Nuclear Corporation (CNNC) and China General Nuclear Power Corporation (CGNPC or CGN)

China’s nuclear sector is dominated by two state-owned giants: CNNC and CGNPC (or simply CGN). Though both are centrally owned, they have distinct mandates and histories and operate under a model of state-guided competition. This “one industry, multiple champions” approach is a hallmark of China’s strategy in critical areas such as energy and infrastructure.

Founded in 1955, CNNC originated as the Second Ministry of Machine Building Industry, in charge of nuclear industry development. In 1988, it was renamed China National Nuclear Corporation (CNNC) and transformed into a state-owned enterprise. CNNC was instrumental in developing China’s first nuclear weapons, submarines, and power reactors. Today, CNNC is considered the more strategic of the two companies, overseeing the full nuclear fuel chain, military nuclear assets, and civilian power plants.

CGN was established in 1994 to focus on civilian nuclear energy and facilitate technology transfer. Originally named China Guangdong Nuclear Power Group, it has since evolved into China General Nuclear Power Corporation (CGNPC), a commercially oriented central State-Owned Enterprise (SOE) known for its operational efficiency and international partnerships.

Nuclear energy in China is treated not merely as a commercial venture but as a strategic imperative. CNNC and CGN are tasked with ensuring energy security, stabilizing electricity supply, and supporting China’s low-carbon transition. While profitability matters—especially to sustain Research and Development (R&D) in advanced technologies—short-term gains are considered secondary to broader national goals.

Projects are typically led by CNNC or CGN, often in joint ventures with local governments and state-owned grid operators. Central government approval signals strong political backing and minimizes future policy risk. Financing is generally provided by the SOEs themselves, local authorities, and state-owned banks offering long-term, low-interest loans underpinned by implicit sovereign guarantees. Land use and licensing processes are streamlined accordingly.

Regulated tariffs—set or approved by the National Development and Reform Commission (NDRC)—offer nuclear operators predictable revenue for 15 to 20 years. In many regions, nuclear plants are given priority grid access over coal-fired plants for environmental reasons. (See China’s Construction Costs and Capabilities for details.)

In 2024, China National Nuclear Power Co., Ltd. (CNNP), CNNC’s primary domestic operating platform, managed 25 reactors with a combined installed capacity of 23.6 GW (gross).59 Including Zhangzhou-1, which entered commercial operation in January 2025, CNNP had a total of 26 reactors online, with a net capacity of 23.2 GW.

CGN Power, the main listed subsidiary of CGN Group, operated 28 commercial reactors, making it a central player in China’s nuclear landscape.60 Those represented 29.6 GW net.

Targets vs. Reality

China has dominated global nuclear power development over the past quarter-century, though its ambitious latest Five-Year Plan targets have proven challenging to meet.

The 10th Five-Year Plan (2001–2005) put forward a policy of “moderate development of nuclear power,” targeting around 8.6 GW gross operating capacity by 2005,61 with 7.1 GW gross achieved in reality. (All Five-Year Plan capacity numbers quoted hereunder are gross gigawatts).

During this period, China connected six new units to the grid—including two French 900-MW reactors at Ling Ao and two Canadian 668-MW CANDU 6 reactors at Qinshan—and completed the development of the CPR-1000, China’s indigenized version of the French M310 900-MW design that would become the workhorse of its early nuclear fleet.

The 11th Five-Year Plan (2006–2010) called on China to pursue “an active development of nuclear power” with a target of 10 GW gross operating by 2010.62 With 10.9 GW gross operating at the end of 2010, that target was slightly over-achieved.

This period saw the construction starts for Westinghouse’s two AP-1000s at Sanmen in 2009 and AREVA’s EPRs at Taishan in 2009–2010, China’s first Gen III projects. Construction commenced on 29 units, most of which were CPR-1000 reactors.

Fukushima’s March 2011 disaster fundamentally reshaped China’s nuclear trajectory during the 12th Five-Year Plan (2011–2015). The government imposed a moratorium on new approvals to conduct comprehensive safety reviews. Existing plants and Gen II reactors under construction had to undergo major upgrades including enhanced flood barriers, backup power system overhauls, and seismic reinforcements.63

When approvals resumed, China adopted a strict “Gen III-only” policy requiring passive safety features and core-catchers. Operational capacity reached just 28.7 GW by 2015 versus a target of 40 GW.64 Nevertheless, the period closed with construction beginning on Fuqing-5 and -6 as well as Fangchenggang-3, China’s first Hualong One reactors, representing its indigenous Gen III technology.

The 13th Five-Year Plan (2016–2020) aimed for 58 GW operational capacity plus 30 GW under construction while establishing the Hualong One as an exportable technology and advancing systems like the high-temperature gas-cooled reactor (HTR-PM) and fast reactors.65

However, domestic capacity reached only 51 GW by 2020 and 17.5 GW under construction constrained by the ongoing inland reactor ban— a controversy unheard of in other nuclear countries limiting nuclear power plant development to the seashore—and extended construction timelines for Generation III units.

In August 2019, the U.S. added CGN to its Entity List,66 citing national security concerns regarding alleged attempts to acquire U.S. technology for military purposes.67 This restricted CGN’s access to certain technologies and affected its international partnerships, including involvement in nuclear projects in the United Kingdom. Later, CNNC was also added to the Entity List.68 The sanctions reinforced China’s focus on self-reliance, accelerating the transition from foreign technologies to the domestically developed Hualong One design.

With an operating capacity of around 61 GW as of mid-2025, the 14th Five-Year Plan (2021–2025)69 target of 70 GW operational capacity is out of reach. According to plans, 4.5 GW are scheduled to come online in 2025, but no new reactor started up in the first half of the year. COVID-19 pandemic disruptions to global supply chains, combined with delays caused by mandatory safety upgrades, have created persistent bottlenecks. First-of-a-kind Hualong One projects saw numerous delays (see Figure 23).

Meanwhile, plans for innovative projects like offshore floating nuclear power platforms appear to have stalled, with 2023 reports suggesting the program may have been suspended over safety and feasibility concerns.70

The Localization of China’s Nuclear Industry

China’s nuclear sector has progressed from technology importer to developing indigenous capabilities over four decades through industrial policy, corporate initiative, and regulatory frameworks. The development of local nuclear technology and independent construction capabilities has remained a consistent priority since the beginning of China’s nuclear power program. Developments include the Hualong One reactor. According to CNNC, for the first such reactor to be commissioned (Fuqing-5), all the key components and parts have been manufactured in China with the project’s localization rate exceeding 85 percent.71 China also developed a localized supply chain for the HTR-PM.

The path to technological independence involved ongoing debates between proponents of importing foreign technology to accelerate learning and advocates for greater self-reliance. CGN was initially tasked with pursuing technology transfer through international partnerships.

Historians Chen Yue and Li Yunyi from the Chinese Academy of Sciences have documented this process by evaluating projects like Daya Bay and Ling Ao, both developed in cooperation with French company EDF.72 The following paragraphs are mainly based on their work.

At Daya Bay-1, constructed from 1987 to 1993, French teams led design, construction, and operations. EDF maintained approval authority for technical decisions and typically preferred French and, to some degree, British suppliers. Although full design blueprints and core technology rights were not transferred, Chinese personnel gained practical experience in construction, installation, and project management by working alongside French and British engineers.

Concurrently, CNNC pursued an independent approach with Qinshan-1, China’s first domestically designed and built nuclear plant. Construction lasted from 1985 to December 1991 with commercial operation only beginning in 1994.

The Ling Ao Phase I project (1997–2002) marked a shift toward localization. While Unit 1 used imported equipment, Unit 2 incorporated domestically manufactured key components such as steam generators, reactor internals, turbine and control rod drive mechanisms—often produced under French supervision.

This approach facilitated knowledge transfer while establishing foundations for local industry. Dongfang Boiler Factory, for example, through this collaboration developed expertise in manufacturing key components including steam generators, boron injectors, safety injection boxes, and voltage regulators. That led Dongfang to become a preferred supplier for subsequent projects like Qinshan-3, Yangjiang, and Ningde.

For Ling Ao Phase II (2005–2011), CGN developed the CPR-1000, a modified version of the French M310 design featuring over 300 technical improvements, including digital safety systems, modernized control rooms, and more efficient turbines. Most design, manufacturing, and construction work was conducted domestically. China also integrated knowledge from other reactor types, including two Canadian CANDU 6 (Qinshan III, 1998–2003) and two Russian VVER-1000 (Tianwan, 1999–2007) designs. Workforce development was supported by onsite vocational training programs for nuclear engineers and technical staff.

Despite progress, localization of certain nuclear island components—particularly in metallurgy, pumps, and instrumentation and control systems—remained challenging. The 2006 Westinghouse agreement to supply four AP-1000 reactors helped address these gaps through included technology transfers.73

Chinese engineers gained access to designs and manufacturing processes for critical systems like the canned motor reactor coolant pump. Companies such as Harbin Electric Corporation and Shanghai Electric studied and adapted these technologies, conducting extensive analysis of over one hundred detailed blueprints and performing fatigue, stress, seismic, and accident simulations.

China’s Construction Costs and Capabilities

China’s nuclear power sector has systematically reduced construction costs while advancing its technological capabilities. According to 2022 investor disclosures from CNNC, the construction cost for Hualong One reactors stood at about CNY16,000/kW (US$20222,375/kW)—significantly lower than the CNY20,000/kW (US$20222,970/kW) for AP-1000 units built in China per the same statement74 and a fraction of the estimated US$15,000/kW for the two AP-1000s completed at Vogtle in the United States.75 The industry continues working toward a target of CNY13,000/kW (US$1,800/kW) for Generation III reactors, which would match China’s official Generation II construction costs.76 There is no independent assessment of these numbers.

Several factors contribute to these apparent significant cost advantages. The sector benefits from a highly skilled workforce cultivated over three decades of continuous nuclear development, according to the analysis of the consultancy affiliated with Chinese state-owned investment holding company China Reform Holdings Corporation.77

Working for state nuclear enterprises like CNNC remains very attractive in China, offering what many view as the “iron rice bowl”—stable, well-paid employment with strong benefits and social prestige. This was evident in CNNC’s 2025 spring recruitment, which reportedly attracted nearly 1.2 million applications from over 425,000 applicants across Chinese universities—every applicant could apply for up to 20 positions—for just 8,000 job openings. CNNC employs around 180,000 people.78

As was the case in the early years of most nuclear programs in the world, the financial model for Chinese nuclear power combines high upfront costs with favorable long-term economics. Plants carry substantial fixed costs including depreciation, financing expenses, and decommissioning funds, but benefit from relatively low variable costs after commissioning.

The National Development and Reform Commission set nuclear electricity prices at CNY430/MWh (US$60/MWh) in 2013—below local coal power benchmarks—while granting nuclear priority grid access that supports high load factors.79 This price level is surprisingly high considering the low investment costs claimed. For comparison, the guaranteed price for retailers of 100 TWh of French nuclear power (ARENH scheme) stands at €42/MWh (US$47.7/MWh) since 2012.80

Analysis from Cinda Securities, a Chinese state-owned asset management bank, assessed the nuclear Levelized Cost of Electricity (LCOE) at around CNY200/MWh (US$202427.8/MWh),81 a fraction of LCOEs in western countries (see Nuclear Power vs. Renewable Energy Deployment). Cinda estimated that projects are expected to achieve around 8 percent for Gen III reactors and 10 percent internal rates of return for Gen II reactors.

China has embraced modular construction techniques, originally pioneered by Bechtel and GE for Boiling Water Reactor (BWR) designs and later adopted for Westinghouse’s AP-1000. This approach involves factory prefabrication of complete components—from primary cooling system parts to steam generators and concrete containment sections—which are then assembled on site. The method is supposed to reduce Hualong One construction timelines to 5–6 years, compared to 8–10 years for conventional builds.

China’s first Hualong One and AP-1000 projects significantly exceeded expected construction times.

Sanmen’s AP-1000s faced long delays—construction times of 111 months for Unit 1 (2009–2018) and 104 months for Unit 2—partially due to post-Fukushima safety modifications and technical issues with the first-of-a-kind passive cooling systems and modular components, particularly the problematic main coolant pump that required redesigning. China’s second AP-1000 at Haiyang took less time—still just under 100 months (2010–2018)—benefiting from lessons learned at Sanmen.

The efficiency gains for Hualong One first appeared more significant. At Fuqing, China’s inaugural Hualong One unit required over 66 months from first concrete in May 2015 to grid connection in November 2020, as engineers made real-time design adjustments and established localized supply chains for critical components like reactor coolant pumps.

A later Hualong One project, Zhangzhou-1 (October 2019–November 2024) was completed in 61 months—5 months faster than Fuqing-5—but appears to be an outlier for the time being. Originally, Zhangzhou-1 was supposed to be an AP-1000 but was later switched to the Chinese design.82 Other Hualong One projects took longer to build: 72 months for Fuqing-6, 84 months for Fangchenggang-3, and over 87 months for Fangchenggang-4 (see Figure 23).

1. Forty Years of Nuclear Constructions in China, 1985–2025

Sources: Various, compiled by WNISR with IAEA-PRIS, various years, 2025

Note: See Table 23 · Chinese Nuclear Reactors in Operation (as of 1 July 2025) in Annex 2 for construction and grid connection dates.

As of 1 July 2025, one of the 14 Hualong One units currently under construction was already exceeding the 61-months benchmark (Taipingling-1 at 65 months). While the average reactor construction times in China remain far below world average (see Figure 24), there appears to be an increasing trend of longer construction times. The reasons remain obscure.

1. Evolution of Reactor Construction Times in China

Sources: WNISR with IAEA-PRIS, 2025

China’s Growing Range of Nuclear Technologies

China’s growing dominance in nuclear technology is evident in its patent filings. A 2024-study by Teva Meyer from the Université de Haute-Alsace analyzed over 56,000 unique nuclear technology patents filed globally between 1946 and 2022.83 The findings reveal China’s acceleration accounting for 48 percent of all nuclear-related patents filed worldwide over the past two decades, with a particularly dramatic rise after 2004.

In 2022 alone, Chinese entities were responsible for 75 percent of global nuclear patent filings. China is particularly active in the development of fourth-generation reactor designs and fuel chain technologies.

Meyer’s analysis suggests multiple motivations behind this patent surge. While some filings may serve performative purposes—signaling technological leadership and independence—the primary driver for patent filings appears to be China’s strategic push to develop indigenous nuclear designs and localize supply chains. This effort aims to reduce dependence on foreign intellectual property, particularly in cases where export controls could potentially weaponize technological dependencies.

Beyond patents, China has made progress in deploying so-called advanced nuclear technologies. On 13 July 2021, China poured first concrete for a demonstration commercial SMR, the ACP100 or “Linglong One”, located on Hainan Island. The project aims at completion by May 2026 after only 58 months of construction.84

According to the builder, several milestones have been achieved ahead of schedule, including the 30 November 2022 installation of nuclear island equipment, which began 75 days early.85 Subsequent milestones included the August 2023 installation of the core reactor module,86 completion of main structural work by February 2024,87 and installation of the main pump in April 2025.88

China’s nuclear developments extend to alternative fuel chains. The 2-MWt liquid fuel thorium-based molten salt experimental reactor (TMSR-LF1) in Gansu Province achieved criticality on 11 October 2023, followed by full capacity operation on 17 June 2024. The reactor does not generate power. The Environmental Impact Report by the Shanghai Institute of Applied Physics (SINAP) released in October 2024 proposes to scale up the experiment and build a 10 MWe thorium-based molten salt reactor nuclear power plant with a maximum thermal output of 60 MWt, with construction proposed to begin in 2025 and the reactor projected to reach full power by 2029.89 This reactor is expected to serve as a research facility to demonstrate system integration and validate key technologies, data, and experience needed for potential future commercial-scale TMSRs.

Non-Proliferation Concerns Around Plutonium Separation and Use

Certain of China’s nuclear developments have raised concerns amongst some non-proliferation experts. Of particular note are the two CFR-600 sodium-cooled fast breeder reactors—their current operational status is unclear—which could theoretically produce significant quantities of weapons-grade plutonium of exceptional quality.

The CFR-600 is a follow-up of the China Experimental Fast Reactor (CEFR) whose operational status is also unclear. The IAEA has simply deleted the CEFR file from its PRIS database. WNISR considers the CEFR in LTO.

While there has been no public confirmation of the CFR600 units connecting to the grid, their progress has attracted international attention due to their potential military applications. The first CFR-600, construction of which began in 2017, was originally scheduled for grid connection in late 2023. Zhang Hui, a Senior Research Associate at Harvard University’s Belfer Center, reports that according to Chinese sources, the first unit has been operating at low capacity since mid-2023.90 Satellite imagery analysis by researchers from the Sasakawa Peace Foundation supports this assessment; comparing images from 23 December 2022 and 16 October 2023 revealed steam emissions from a ventilation stack in the later image, suggesting operational activity.91 However, as of mid-2025, its operational status remains unclear, and it is not known whether it is connected to the grid. The second unit is planned for operation in 2026.

According to a 2018 bilateral agreement, Russia will supply the fuel for the first unit for its entire lifetime.92 Moreover, highly enriched uranium fuel, delivered by Russia under a 2019-contract for the first seven years of operation, would be subject to peaceful use restrictions according to Russian export regulations and a China-Russia nuclear cooperation agreement.93 A 2019 Rosatom statement claims that the “CFR-600 will first use mixed [uranium-plutonium] oxide (MOX) fuel” and that “both the operating CEFR and the under-construction CFR-600 are designed to use MOX fuel that is not produced in China.”94 It remains unclear with what kind of fuel the CFR-600 has or will be loaded and what the fuel management will look like in the longer term.

Further raising non-proliferation questions is China’s reprocessing infrastructure. Satellite images analyzed by Harvard’s Zhang Hui suggest CNNC may have begun construction on a third reprocessing plant, adjacent to two existing facilities, at its Nuclear Technology Industrial Park in Jinta, Gansu.95

China began construction of the first two reprocessing plants (each with 200 tons/year capacity)96 in 2015 and 2020, respectively. By late 2024, major construction appeared complete, with commercial procurement documents indicating equipment installation since 2023. Zhang estimated at the time that these plants could begin operation in 2025, five years earlier than expected.

The purpose of the potential third plant remains unclear due to limited public information. Zhang suggests it could either support the first two plants’ reprocessing activities or represent a modular expansion approach. Notably, these facilities are located at the same site as a demonstration MOX fuel fabrication line (20 tons/year capacity), expected to be commissioned in 2025 to supply fuel for China’s second CFR-600.

Since 1983, China has followed a plutonium separation and use policy goal aimed at managing spent fuel and reducing uranium consumption. The State Council approved a pilot reprocessing plant at Jiuquan in 1986, which conducted its first test run in 2010.97

The dual-use potential of the plutonium fuel system has drawn some international concern. U.S. Defense reports have warned about possible military applications,98 while China’s transparency has decreased notably, e.g. it stopped reporting separated plutonium stocks to the IAEA in 2017.

The High-Temperature Reactor Program

The High-Temperature Reactor Pebble-bed Modules (HTR-PM) project at Shidao Bay represents the culmination of nearly four decades of research by Tsinghua University’s Institute of Nuclear and New Energy Technology (INET), beginning with a 10-MWt test reactor in 1984.

The current demonstration plant—originally with a Reference Unit Power of 200 MWe, reduced for unknown reasons to 150 MWe—developed through collaboration between Tsinghua University, China Huaneng, and CNNC, features two reactor modules driving one common steam turbine. Construction began with first concrete poured on 9 December 2012, leading to grid connection of the first module on 20 December 2021. The exact date of the second module’s grid connection has not been published, but it is likely to have taken place during the year 2022. This means that construction time was around 10 years, twice as long as planned at construction start. After the completion of a 168-hour demonstration run, the plant entered commercial operation on 6 December 2023.99

While the HTR-PM designers claim several advantages over other technologies, the technology faces challenges. Economic viability remains to be proven, with estimated power generation costs approximately 20 percent higher than conventional pressurized water reactors due to lower power density and complex fuel fabrication.100 The pebble-bed design produces bulky graphite waste, and the fuel handling systems require optimization. Current operating temperatures (~685°C) fall short of the 750°C design target, limiting efficiency for applications like hydrogen production. Regulatory hurdles also persist due to the novel design’s lack of precedent.

The design is based on a German concept that was an industrial disaster. The 300-MW THTR (Thorium High Temperature Reactor) in Hamm-Uentrop took 14 years to build and was less than three years on the grid before being closed.101 Countless technical problems had convinced the owners to seal the fate of the plant. The reactor is the only one in Germany that is still in long-term enclosure, and 37 years after its closure, decommissioning has not begun (see Germany in Decommissioning Status Report).

The Chinese technology promoters plan to address the challenges of the first HTR-PM through the HTR-PM600 project, designed to expand the modular concept to six 100-MW modules driving a single turbine. This demonstration project, to be paired with Hualong One reactors in Xuwei, Jiangsu Province, is planned to supply industrial steam to nearby petrochemical facilities in Lianyungang.102

Nuclear and Renewable Energy

In China, nuclear and renewable energy are both seen as vital for China’s dual goals of carbon reduction and energy security while maintaining grid stability. They are generally not regarded as competing energy sources but as inhabiting complementary roles, aided by geographic separation: nuclear plants are located along the populous, coastal provinces, where they benefit from abundant cooling water and meet steady demand from electricity intensive sectors such as manufacturing or textiles. In contrast, large-scale wind and solar projects are concentrated in remote inland areas like Inner Mongolia, Xinjiang, and the Gobi Desert, where space and resources are plentiful.103

This distribution reduces direct competition with nuclear delivering baseload power to coastal cities, while wind and solar transmit power over long distances through ultra-high-voltage (UHV) lines to industrial centers. However, distributed solar has been reaching a very significant share of new additions, with over 44 percent in 2023 and even 58.5 percent in 2022. In the first half of 2024, the share was still at 44 percent.104 By the first quarter of 2025, the share of distributed solar jumped to 60 percent—36 GW of a total of 60 GW—according to some estimates.105 (See section on China in Nuclear Power vs. Renewable Energy Deployment).

China’s power grid follows a semi-regulated dispatch system that prioritizes low-carbon sources. Renewables have top priority due to policy mandates and negligible marginal costs, followed by nuclear power, which operates as high-priority baseload with minimal flexibility. Hydropower is used for peak demand, while coal and gas act as buffer sources that ramp up or down as needed.

But given China’s enormous size, there are regions where nuclear competes with renewable energy for dispatch priority, for example, Guangdong, which hosts several nuclear facilities. Its nuclear fleet, dominated by large Generation II PWRs, was designed primarily for stable baseload generation. These reactors typically operate at high load factors—close to 90 percent—and are not suited for flexible load-following, meaning they cannot easily ramp output up or down in response to fluctuations in electricity demand or renewable energy supply (see chapter on Challenges of Integrating Nuclear Power into the Energy System).

During periods of low demand or high renewable output, the grid can become congested. In 2016 to 2017, this caused some nuclear power plants to reduce output or even shut down, resulting in higher system costs, wasted nuclear fuel, and unnecessary subsidies for unused renewable power.106

The authors of a recent paper on the “dynamics of power curtailment in China” highlight “the necessity of mitigating grid congestion, enhancing inter-provincial transmission, and adopting market-driven pricing strategies to maintain the long-term cost-competitiveness of renewable energy.” And they stress the role of storage: “As the price of storage drops and its role in the system becomes more critical due to the rise of variable renewables, its deployment is projected to expand.”107 Indeed, in 2024, China increased new grid-connected battery capacity by 130 percent to reach a total of 74 GW/168 GWh.108

France Focus

Overview

Nuclear generation increased by 41.3 TWh (+12.9 percent) to 361.7 TWh, representing 67.1 percent of total French power generation in 2024.109 Between 2005 and 2015, the norm was French nuclear production of 400 TWh, peaking at 430 TWh in 2005. Meanwhile, nuclear power generation declined in the rest of the EU27 (see Figure 28).

In December 2024, national utility Électricité de France (EDF) connected Flamanville-3 (FL3), its first European Pressurized Water Reactor (EPR) in the country, to the national grid, 17 years after construction start, 12 years later than planned, and 25 years after the last unit, Civaux-2, was added to the French program. So far, FL3 has not been performing as planned, and the new Authority for Nuclear Safety and Radiation Protection (ASNR) noted: “About fifty significant safety events were however declared by the licensee between commissioning and the end of 2024, a rate that is significantly higher than expected, even for a new reactor.”110

In June 2023, the National Assembly passed legislation for the “acceleration of procedures for the construction of new nuclear facilities near existing nuclear sites and for the operation of existing facilities” (see France Focus in WNISR2023).111 While these measures can cut some red tape, they are unlikely to significantly ease the phenomenal industrial challenges.

In February 2022, the French President announced a plan to build six units of a new design called EPR2, with first startup targeted for 2035. Moreover, he stated that the option of building eight additional units by 2050 should be studied.112

The EPR2 still does not exist on the drawing board; no detailed design is available yet. The government estimated in an October 2021 internal note that, if everything goes well, the first EPR2 could start up by 2039–2040. In case unexpected industrial difficulties occur—as they did in the past and currently do—it could take until 2043 to commission the first EPR2, the project review stated.113

It is only on 22 July 2024 that the beginning of the development from basic to detailed design was accepted during a “Review Committee” meeting.114 EDF plans to provide an updated cost-assessment to the government by the end of 2025, and a Final Investment Decision (FID) is anticipated for the end of 2026.

In January 2025, the French Court of Accounts reported that “the projected profitability of the EPR2 program remains unknown at this stage, especially since the financing terms for this program have still not been determined.”115 The lack of an updated cost assessment, and thus the absence of a financing package and an FID, did not prevent EDF from ordering large components like reactor pressure vessels and steam generators116 or from starting site preparations.

On 1 January 2025, the law of 21 May 2024 took effect that laid the ground for the absorption of the Institute for Radiation Protection and Nuclear Safety (IRSN) by the Nuclear Safety Authority (ASN) and the creation of the Nuclear Safety and Radiation Protection Authority (ASNR). The legislation was implemented through a government decree issued on 3 December 2024117 against a remarkably broad opposition, from conservative and ultra-pro-nuclear MPs to anti-nuclear Greenpeace and IRSN’s ‘intersyndicale’, a coalition of trade unions represented at the institute and its Social and Economic Committee (CSE).118 The move constitutes a roll-back of a forty-year long history of the progressive establishment of a Technical Support Organization more independent of the nuclear industry, the state, and the safety authority. IRSN was originally a department of the Atomic Energy Commission (CEA); it was given an independent status and later merged with the national radiation protection office.

On 17 March 2025, President Macron chaired a session of the Nuclear Policy Council (CPN) that since 2022 “defines the broad guidelines for national nuclear policy”.119 Thus in the self-understanding of the presidency, it is not the parliament, but the CPN chaired by the President of the Republic that provides the framework orientation of nuclear policy for the long term.

The CPN took note of the status of the EPR2 program aiming at a first startup by 2038 and asked EDF “to step up its efforts to control costs and the schedule and to present a binding estimate of costs and deadlines by the end of the year.” The CPN has examined the financing scheme of the program based on a government loan covering at least half of the construction costs and a contract for difference at a maximum price of €2024100/MWh (US$2024108.7/MWh). The support scheme needs to be approved by the European Commission.

The CPN also “confirmed the continued investments” into Orano’s backend strategy at the La Hague site, including the construction of additional spent-fuel storage capacities and a new uranium-plutonium mixed oxide (MOX) fuel fabrication plant—both to be commissioned by 2040. Further, the CPN “approved the principle of financing this program mainly born by EDF, as the future customer of these facilities…”120 None of these decisions has undergone specific parliamentary scrutiny, but they have triggered significant local opposition.121

No Reactor Under Construction in France—Again

Until the closure of the two oldest French units at Fessenheim in spring 2020, the French nuclear fleet had remained almost stable for 20 years, except for the closure of the 250-MW fast breeder Phénix in 2009 or the two units in Long-Term Outage (LTO) between 2015 and 2017. Another unit was in LTO in the period 2021–2023 (see Figure 25). Penly-1, subject to the stress-corrosion cracking issue, was offline between 2 October 2021 and 13 July 2023.122 No unit was offline for the entire year of 2024.

1. Operating Fleet and Capacity in France

Sources: WNISR with IAEA-PRIS, 2025

Flamanville-3 was the first new reactor to start up since Civaux-2 was connected to the French grid 25 years earlier in December 1999.

The first and only Pressurized Water Reactor (PWR) closed prior to Fessenheim was the 300-MW Chooz-A reactor, which was retired in 1991. The other closures were those of eight first-generation natural-uranium gas-graphite reactors, two fast breeder reactors, and a small prototype heavy water reactor (see Figure 26).

1. Startups and Closures in France

Sources: WNISR with IAEA-PRIS, 2025

Notes: PWR: Pressurized Water Reactor; GCR: Gas-Cooled Reactor; HWGCR: Heavy Water Gas Cooled Reactor; FBR: Fast Breeder Reactor.

In 2024, the 57-reactor fleet123 produced 361.7 TWh, an increase of 41.3 TWh (+12.9 percent) over the previous year; the production remained at a level below that of the pre-COVID-19 year 2019 and the ninth year in a row below 400 TWh.

In 2005, nuclear generation peaked at 430 TWh. After the construction program was completed in 1999, it took the fleet five years to build up to that maximum generation, and with a quasi-stable installed nuclear capacity between late 1999 and early 2020, performance plunged after 2015 (see Figure 27).

In 2024, nuclear plants provided 67.1 percent (+2.1 percentage points) of the country’s electricity. The nuclear share peaked in 2005 at 78.3 percent. As of mid-year 2024, EDF estimated production in the 335–365 TWh range for 2025 and 2026.124 In January 2025, the expected production was reevaluated at 350–370 for the years 2025–2027 (see Figure 27 and Figure 28).125

1. Nuclear Electricity Production vs. Installed Capacity in France

Sources: RTE, 2000–2025; and EDF, 2025

Note: In Figure 27, reactors in LTO are counted in the “installed capacity”.

The year 2024 saw record new additions of solar capacity (+5 GW or +26 percent) to reach a total of 24.3 GW as well as a record solar generation of 24.8 TWh, a 10 percent year-on-year increase.126 In the first half of 2025, another 2.8 GW were connected to the grid. There are now close to 1.2 million solar power facilities in metropolitan France,127 with 63 percent (or 18 percent of the installed PV capacity) practicing partial or total self-consumption of the power they produce as of 2024. This last segment is growing fast and the installed capacity operated in some form of self-consumption has increased by 72 percent between the end of 2023 and end of 2024.128

1. Nuclear Electricity Production vs. Nuclear Share in France

Sources: RTE, 2000–2025; and EDF, 2025

Since 2020, in every single year monthly nuclear production in the first half-year remained below the average level of the period 2012–2019. In 2024, production improved significantly especially in the crucial fourth quarter (see Figure 29).

Electricity represented 26.6 percent of final energy consumption in France in 2024. As nuclear plants provided 67.1 percent of electricity, they covered 17.9 percent of final energy. Fossil fuels covered the largest share at 57.3 percent, with oil at 39.3 percent and natural gas at 17.5 percent (coal <1 percent).129

1. Monthly Nuclear Electricity Generation, 2012–mid-2025

Sources: RTE and EDF, 2021–2025130

Nuclear Unavailability Review 2024131

In 2024, there were 5,539 reactor-days with zero-production per reactor—a significant drop of 22 percent compared to 2023 and close to 35 percent less than in 2022, but still an average of 99 days or over three months. This includes neither load following nor other operational situations with reduced but above-zero output. The 2024 number is similar to the average 96 days outage days per reactor in pre-COVID year 2019. All fifty-six reactors (not including FL3) were subject to outages lasting from nine to 258 days (see Figure 31). No reactor has been offline the whole year.

Table 4 illustrates that while the declared “planned” outage-days dropped significantly in 2023 and 2024, the declared “forced” outages at 342 days remained on a high level in 2024, only second behind 2023 in the past six years.

1. Total Unavailability at French Nuclear Reactors, 2019–2024 (in Reactor-Days)

	Declared Type of Unavailability
	“Planned”	Forced	Total	Average per Reactor
2019	5,273	316	5,588	96
2020	6,179	286	6,465	115
2021	5,639	172	5,811	104
2022	8,287	278	8,515	152
2023	6,704	399	7,103	127
2024	5,197	342	5,539	99

Sources: RTE and EDF REMIT Data, 2019–2025

Note: The year 2020 does not include Fessenheim-1 and -2 that were closed during the year, and the year 2024 does not include Flamanville-3 that was briefly connected to the grid in December 2024.

1. Reactor Outages in France in 2024

Sources: compiled by WNISR with RTE and EDF REMIT Data, 2021–2025

Notes: For each day in the year, this graph shows the total number of reactors offline, not necessarily simultaneously, as all unavailabilities do not overlap, but on the same day. This figure does not take into account Flamanville-3 that was briefly connected to the grid in December 2024.

The unavailability analysis for the year 2024 in Figure 30 further shows that outage numbers and durations are down in all categories, sometimes by large margins:

• During the whole year at least 7 units and up to 24 were down during the same day.
• On 59 days (16 percent of the year)—a remarkable decrease from 252 days (69 percent of the year) in 2023—19 or more units were shut down for at least part of the day.
• At least six reactors were down (zero capacity) simultaneously on any day of the year.
• At least 20 reactors were offline simultaneously during the equivalent of 21 days, down from 192 days in 2023.

According to EDF’s classification of “planned” and “forced” unavailabilities, in 2024:

• 16 reactors did not experience any “forced” outage.
• At five units “forced” outages lasted less than one day.
• At 28 units cumulated “forced” outage duration represented between one and 10 days.
• At 7 reactors “forced” outage cumulated between 10 and 64 days over the year (see Figure 31).

1. Forced and “Planned” Unavailability of Nuclear Reactors in France in 2024

Sources: compiled by WNISR with RTE and EDF REMIT Data, 2021–2025

Notes: This graph only compiles outages at zero power, thus excluding all other operational periods with reduced capacity >0 MW. Impact of unavailabilities on power production is therefore significantly larger. Flamanville-3, briefly connected to the grid in December 2024, is not considered here.

“Planned” and “Forced” unavailabilities as declared by EDF.

However, EDF’s declaration of “planned” vs. “forced” outages is highly misleading. EDF considers an outage as “planned”, whatever the number and length of extensions (or, in rare cases, reductions) of its total duration, if the outage was first declared as “planned”.

Detailed WNISR analysis for earlier years shows a different picture.

1. Unavailability of French Nuclear Reactors, 2020–2024

Sources: compiled by WNISR with RTE and EDF REMIT Data, 2019–2025

Notes: The categorization as “planned” or “forced” outage follows EDF’s classification. However, it does not reflect reality, as EDF keeps a “planned” outage in that category even if it lasts much longer than “planned”.

The cumulated outage analysis over the five years 2020–2024 reveals the following (see Figure 32):

• Five reactors were down half of the time or more (Flamanville-1 and-2, Chooz-1 and -2, and Penly-1);
• 32 reactors were generating zero power for 30 percent of the time, that is 109 days, or more per year on average.
• 45 reactors were off-grid for at least one quarter of the time, in other words, they did not generate any power for the equivalent of one in four years.

Stress Corrosion Cracking and Thermal Fatigue

Severe stress corrosion cracking had been first identified in late 2021 on the safety injection systems of the four largest and most recent French reactors at Chooz and Civaux.132 Later, additional impacted reactors were identified and a program of pre-emptive replacement of particularly sensitive piping sections was decided for the “P’4” reactor series. While apparently rare so far, the phenomenon has also been identified on other 1300-MW and some 900-MW reactors. EDF decided to inspect its entire reactor fleet by the end of 2025 and claimed that, as of mid-2024, already 50 of its 56 units had been “controlled and treated”.133

In February 2023, an additional issue was identified during destructive examination at Penly-1. Close to a weld of a line of the safety injection system that had been repaired during construction of the plant, a 15.5 cm long (about one quarter of the circumference) and up to 2.3 cm deep crack (for a 2.7 cm thick tube) was identified. The origin has been determined as thermal fatigue rather than stress corrosion cracking. This discovery meant that an extensive inspection program of all repaired welds had to be added to the stress corrosion cracking investigations. According to the plan, 90 percent of the repaired welds in the safety injection and shutdown cooling systems of the entire reactor fleet were to be inspected by the end of 2024 and the remaining ones in 2025.134

In June 2025, EDF discovered new crack indications on two welds of the Civaux-2 shutdown cooling system; one was determined as stress corrosion cracking, the other as thermal fatigue. EDF is confident that “the control tools and repair methods developed by EDF since 2022 have been mastered and industrialized.”135

As of mid-2025, EDF displayed a message on its website saying:

Taking into account key learnings from controls and repairs undertaken in 2023 and 2024 on reactors linked to stress corrosion cracking, EDF considers that, in 2025, one planned outage out of six (…) may be extended by an average of 30 days.136

EDF indicates that the following reactors are concerned: Blayais-4, Cruas-1, Dampierre-1, Gravelines-1, St.-Laurent B1, Tricastin-2, Belleville-1, Belleville-2, Cattenom-3, Flamanville-2, Golfech-2, and Chooz-B1.

ASNR in the ASN Annual Report 2024 appears significantly more prudent as to EDF’s mastering of the stress corrosion phenomenon:

On the occasion of the next periodic safety reviews, the lessons learned will need to be incorporated by EDF into its maintenance programmes. ASN in particular considers that the hypotheses concerning the lack of susceptibility to degradation mechanisms, adopted for certain zones which are not monitored by a preventive maintenance programme, must be backed up by a programme of additional investigations.137

Decennial Inspections and Lifetime Extensions

By mid-2025, the average age of the 57 nuclear power reactors exceeded 39 years (see Figure 33); 52 operating units are now over 31 years old of which 27 are over 41 years. Lifetime extension beyond 40 years requires significant additional upgrading. Also, relicensing is subject to public inquiries reactor by reactor.

1. Age Distribution of the French Nuclear Fleet

Sources: WNISR with IAEA-PRIS, 2025

EDF will likely seek lifetime extensions beyond the 4th Decennial Safety Review (VD4) for most, if not all, of its remaining reactors. President Macron in his February 2022 programmatic speech made it clear that the government has no intention of closing reactors anymore. He stated: “While the first extensions beyond 40 years have been implemented successfully since 2017, I’m asking EDF to examine the conditions of the [lifetime] extensions beyond 50 years, in conjunction with the nuclear safety authority.”138

The first reactor to undergo the VD4 was Tricastin-1 in 2019, followed by Bugey-2 and -4 in 2020, and Tricastin-2 in 2021. Dampierre-1, Bugey-5, and Gravelines-1 started in 2021… until the COVID-19 pandemic further disrupted the safety review schedule.139 Until 1 July 2025, 22 units had undergone their VD4, and a further three were underway (see Table 5).140

While ASN judged the VD4 premiere on Tricastin-1 “satisfactory”, it questioned whether EDF’s engineering resources were sufficient to carry out similar extensive reviews simultaneously at several sites.141 Beyond the human resource issue, the experience raised the question of affordability. EDF had scheduled an outage of 180 days for Tricastin-1 in 2019, which was first extended by 25 days to 205 days. Including further, unrelated unavailabilities, the reactor was ultimately in full outage for two thirds of that year (232 days). Many of the following VD4 exercises also saw significant delays between expected and real durations, e.g. Bugey-2 +118 percent compared to plan, Chinon B1 +76 percent (also holds the record outage of 467 days), or Blayais-1 +75 percent (see Table 5).

Of 22 VD4 outages completed until mid-2025, 16 were expanded, while six were completed faster than anticipated. While the delays total an extra 1,444 unplanned outage days, the six earlier completions total only 87 saved outage days, thus a net cumulated unplanned loss of 1,357 production days.

EDF expects the VD4 outages to last six months, much longer than the average of three to four months experienced through VD2 and VD3 outages. In 2023, the Chief Technical Officer of EDF Group and CEO of EDF R&D, Bernard Salha, told French Parliament that the work volume of a VD4 was five times larger than that of a VD3. He also said investments into the operating fleet had doubled over the previous decade.142

1. Fourth Decennial Visits of French 900-MW Reactors, 2019–2025

Reactor	Capacity	Grid Connection	VD4 Outage	Expected Duration*	Total Actual Duration
Tricastin-1	915	31 May 1980	01/06/19–23/12/19	180	205
Bugey-2	910	10 May 1978	18/01/20–15/02/21	181	395
Bugey-4	880	8 March 1979	22/11/20–24/06/21	226	214
Dampierre-1	890	23 March 1980	19/06/21–05/02/22	170	231
Tricastin-2	915	7 August 1980	06/02/21–26/07/21	180	170
Bugey-5	880	31 July 1979	31/07/21–21/04/22	189	265
Gravelines-1	910	13 March 1980	14/08/21–11/04/22	188	240
Tricastin-3	915	10 February 1981	12/03/22–21/11/22	171	254
Gravelines-3	910	12 December 1980	23/03/22–22/12/22	191	275
Dampierre-2	890	10 December 1980	27/04/22–31/12/22	171	248
Blayais-1	910	12 June 1981	31/07/22–19/06/23	185	323
Saint-Laurent B-2	915	1 June 1981	20/01/23–20/11/23	223	304
Chinon B-1	905	30 November 1982	07/02/23–19/05/24	265	467
Gravelines-2	910	26 August 1980	10/06/23–07/03/24	197	272
Blayais-2	910	17 July 1982	24/06/23–31/03/24	182**	281
Dampierre-3	890	30 January 1981	23/09/23–2/03/24	170	161
Bugey-3	910	21 September 1978	11/11/23–28/08/24	177	303
Tricastin-4	915	12 June 1981	19/01/24–16/07/24	194	179
Gravelines-4	910	14 June 1981	20/01/24–07/09/24	195	231
Blayais-3	910	17 August 1983	08/06/24–26/12/24	191	201
Dampierre-4	890	30 January 1981	12/07/24–21/12/24	188	162
Cruas-3	915	14 May 1984	04/08/24–10/03/25	232	218
St-Laurent B-1	915	21 January 1981	31/01/25–14/08/25	178	195
Blayais-1	910	12 June 1981	05/04/25–09/10/25	187
Cruas-1	915	29 April 1983	14/06/25–01/12/25	170

Sources: compiled by WNISR with EDF REMIT-Data, 2019–2025143

Notes: The expected duration is based on outage dates in use as of outage start or within the few days after the reactor is disconnected from the grid.

For ongoing decennial visits, end of outage date is the date in use as of 20 August 2025 and can vary from the original date.

* Expected duration as of outage start - ** Original duration144

On 23 February 2021, ASN issued detailed generic requirements for plant life extension.145 The key aspects of ASN’s decision were not the five short administrative articles but the two annexes setting the technical conditions and the timetable for work to be carried out. The challenge for operator EDF is high, as ASN outlined at the time:

Over the coming five years, the nuclear sector will have to cope with a significant increase in the volume of work that is absolutely essential to ensuring the safety of the facilities in operation.

Starting in 2021, four to five of EDF’s 900 Megawatts electric (MWe) reactors will undergo major work as a result of their fourth ten-yearly outages. (…)

All of this work will significantly increase the industrial workload of the sector, with particular attention required in certain segments that are under strain, such as mechanical and engineering, at both the licensees and the contractors.146

In fact, in 2024 seven 10-year inspections took place, five VD4 for 900-MW reactors and two VD3 for 1300-MW units. Between 2025 and 2028, four (2025, 2027, 2028) to five (2026) 10-year inspections are scheduled annually. In 2029, a VD5 of a 900-MW unit is planned for the first time, alongside four VD4s of 1300-MW reactors and the first VD3 of a 1450-MW unit. Between 2029 and 2034, each year six or seven VDs are scheduled.147

ASN has shown remarkable tolerance for extended timescales of refurbishments and upgrades in the past; many of the post-Fukushima measures have not yet been implemented 14 years after the events, for example. According to some estimates, the completion of the work program could take until 2039.148

Additionally, the implementation of work to be carried out as part of the lifetime extension beyond 40 years stretches over 15 years until 2036, when the last 900-MW reactor is supposed to be upgraded: Chinon B-4, connected to the grid in 1987, gets the 15-year delay to implement 15 of a total of 37 measures. By then, the unit will have operated for 49 years. This is just one example, and it is the newest of the operating 900-MW reactors. The ASN has accepted similar timescales for all 32 of the 900-MW units. The French Nuclear Safety Authorities have proven flexible, and—considering the dire state of the reactor fleet—pressure for even more flexibility might increase in the future.

On 13 October 2023, EDF applied for permission to delay many of the required upgrades of the 32-unit fleet of 900-MW units by years, “given the difficulties of meeting them”.149 ASN, on 19 December 2023, modified its February 2021 publication of generic requirements, and delayed target dates for 31 of the 32 units for at least one but up to 14 upgrading work-packages (of a total of 36 per reactor) for periods of one to five years.150

In the framework of the legally required upgrading work and documentary deliveries, EDF had to meet 107 deadlines in 2024. EDF claims having met all of them.151 Apparently, EDF has reacted to ASN’s alert in 2021 on the immense workload ahead (see above), hired additional staff, and adjusted the project management. EDF concludes its 2024 review on a positive note: “The analysis developed in this report does not identify any alerts concerning a risk of non-compliance with future deadlines for requirements.”152

There are many more target dates coming up over the next two years. In 2025, 105 deadlines are expected with 30 relating to modifications to facilities while 75 relate to documentary deliverables or modifications to the plant’s operating specifications. In 2026, 82 prescription deadlines are expected with 42 relating to work on facilities and 40 relative to documentary deliverables or modifications to the facility operating reference system.153

Financial Issues

Operating and maintenance costs of the ageing fleet of reactors have significantly increased over the past decade (see also previous WNISR editions), but whatever the uncertainties over various cost estimates might be, there is little doubt that the additional costs for refurbishment and upgrades in view of lifetime extensions remain below any cost estimate for newbuild.

Outages that systematically exceed planned timeframes are particularly costly. EDF’s net financial debt increased by about €10 billion (US$202310.6 billion) over the period 2019–2021 to a total of €43 billion (US$202151 billion) as of the end of 2021.154 In 2022 alone, net debt jumped by €21.5 billion (US$202222.6 billion) to €64.5 billion (US$202267.9 billion) at the end of the year.155

In 2023 and 2024, EDF profited from increased nuclear and hydro power output as well as good years for renewables’ generation. In 2023, EDF managed to go from a record loss to a €10 billion (US$202310.8 billion) profit and was able to reduce its debt load by an equivalent amount to €54.4 billion (US$202358.8 billion).156 EDF’s 2024 results show a stabilization of the debt load at €54.3 billion (US$202459 billion). During the year, sales were down 15.7 percent “as prices fell in the countries where the Group does business”, and EDF’s extreme financial sensitivity to market prices was revealed by the fact that, “The downturn in sales prices had an estimated impact of -€18.5 bn [US$202420.1 billion].”157

Rapidly changing market situations leading to lower average prices and, increasingly often, to very low or even negative prices on the spot market, have a significant impact on the management of France’s nuclear fleet and on financial results. In the first half of 2025, EDF saw 769 hours or 18 percent of the time with prices below €10/MWh (US$11.4/MWh), and net income was down 21.2 percent compared to the first half of 2024. At the same time, modulation of nuclear reactors increased by 16 percent compared to the same period in 2024.158

The 20-Year-Long Flamanville-3 EPR Saga—No End in Sight?

The 2005 decision to construct Flamanville-3 (FL3) was mainly motivated by the industry’s attempt to confront the serious problem of maintaining nuclear competence. Fifteen years later, ASN still drew attention to the “need to reinforce skills, professional rigorousness and quality within the nuclear sector.”159 EDF justified its refusal to provide its estimates for FL3’s expected profitability for the Court of Accounts’ January 2025 report by stating in particular: “The project’s main issues at stake were to maintain the skills of the French nuclear industry and to prepare for the deployment of EPR technology in France and around the world.”160

In December 2007, EDF started construction on Flamanville-3 (FL3) with a scheduled startup date of 2012. The project has been plagued with design issues and quality-control problems, including basic concrete and welding difficulties similar to those at the Olkiluoto (OL3) project in Finland, which started construction two and a half years earlier and was connected to the grid only in March 2022 (see earlier WNISR editions). These problems never went away.

Following numerous revisions of the original €3.32007 billion (US$20074.5 billion) cost estimate, in January 2022 EDF estimated the overnight costs at €201512.7 billion (US$201514 billion).161 In 2020, the French Court of Accounts estimated the total cost, including financing and other associated costs, at €201519.1 billion (US$201521 billion).162 The court estimated that the cost of electricity from FL3 would be €2015110–120/MWh (US$2015122–133/MWh).

These estimates have been updated by EDF and corrected by the court in its January 2025 report.163 EDF’s latest estimate of November 2023 indicates overnight costs164 of €18.1 billion (US$16.9 billion) and a total, including financing and other costs, of €22.6 billion (US$24.4 billion) that the court upped to €23.7 billion (US$25.6 billion), a staggering sixfold increase over the original cost estimate (inflated) of €4 billion (US$4.3 billion) (see Figure 34).

1. Construction Times and Project Cost Estimates for Flamanville-3

Source: Cour des comptes, 2025

The court regrets FL3’s “poor projected profitability”. To achieve a rate of return of 4 percent (real term) FL3’s power would need to sell at €138/MWh (US$149/MWh), provided the plant reaches a load factor of 75 percent—the average lifetime load factor of the four latest French reactors at Chooz and Civaux is around 70 percent. The court further states that “for sales prices below €90/MWh [US$97/MWh], it seems difficult to envisage a rate of return reaching 2 percent.” The average spot market price for base load power in the first half of 2025 stood at €66.70/MWh (US$75.7/MWh) in France.165

The fuel issue that struck the Taishan EPRs (in China) and kept Unit 1 off-grid for over one year had consequences for FL3. EDF decided to refabricate 64 of the 241 fuel assemblies that had already been produced. These were approved by ASN and delivered to the site. Fuel loading was finally completed in May 2024, and subsequently, EDF representatives repeatedly stated the reactor startup would be “imminent” with grid connection happening “a few weeks later”.

FL3’s first criticality took place on 3 September 2024, and it was finally connected to the grid a few months later on 21 December 2024, 17 years after construction start and 12 years later than planned.

The startup procedures appear as laborious as the construction and, according to ASNR, “the first months of operation showed that EDF needs to reinforce its oversight and control of operational activities.”166

The EPR2 Project

As early as August 2023, EDF had applied for a building permit for the first pair of the “sixpack” to be built at the Penly site. The ASN states in its Annual Report 2024 that the application “is being examined and should reach a conclusion in 2027.” Site preparation has begun, nevertheless. The other pre-selected sites for a pair of EPR2 units are Bugey and Gravelines, and EDF is in the course of filing all administrative applications to implement these projects. Public debates, part of the procedure, already took place for all three sites: for Penly between 27 October 2022 and 27 February 2023, for Gravelines 17 September 2024 to 17 January 2025, and for Bugey 28 January to 15 May 2025.167

Michel Badré, former president of the Environmental Authority and of the Special National Commission for Public Debate (CPDP), in charge of the public debate on the EPR2 projects in Penly, summed up his sobering conclusion from the procedure in an October 2024 presentation:168

Two criteria for assessing the usefulness of the debate:

- Did it provide the public with adequate information? To date, no.

- Did it allow the public to participate in the decision-making process? No.

The Court of Accounts concludes its already mentioned 2025 report on the EPR by stating:

This analysis shows that, although the French nuclear industry has begun to organize itself to implement the strategy it set out in 2022, it is far from ready and still has many challenges to overcome, some of which are concerning.169

The court warns that “the accumulation of risks and constraints could lead to the failure of the EPR2 program.” Therefore, it has issued two key recommendations:

- Postpone the final investment decision on the EPR2 program until financing has been secured and detailed design studies have progressed in line with the target schedule for the first nuclear concrete pour [the official construction start].

- Ensure that any new international project in the nuclear field generates quantifiable gains and does not delay the EPR2 program schedule in France.

The analysis is based on some raw numbers. At the end of 2023, EDF’s overnight cost estimate for the first three pairs of EPR2s was reevaluated at just under €80 billion (US$202386.4 billion), 30 percent higher than the previous estimate from only one year earlier. In the underlying assumptions, the construction start of the first Penly unit was delayed from end of 2027 to September 2028 and its startup from September 2036 to July 2038—that is already three years later than assumed at the beginning of the official EPR2 planning.

The two nuclear companies EDF and Orano/AREVA alone have absorbed €23.7 billion (US$202425.8 billion) or 86 percent of all public cash injections to industrial companies by the State’s Holdings Agency (APE).

The amounts of money invested into the EPR2 project prior to any construction start are quite extraordinary. The EPR2 Project Review Committee stated in its November 2024 report: “It should be noted that €3 billion [US$20243.3 billion] will already have been spent by EDF by the end of 2024, while an additional €2 billion [US$20242.2 billion] is planned for 2025 and €3 billion [US$20243.3 billion] for 2026, while the final investment decision (FID) cannot be taken before mid-2026, in a timetable that remains optimistic.”170 A total of €8 billion (US$20248.7 billion) is to be spent prior to the FID. Reportedly, EDF later decided to reduce expenditures in 2025 from €2 billion to €1.1–1.3 billion,171 which does not change the order of magnitude of the extraordinary financial engagement prior to the official decision to go ahead.

It is obvious that the French state will have to shoulder much of the EPR2 investment burden. The engagement in favor of the nuclear sector is obviously not new. According to the Court of Accounts, the two nuclear companies EDF and Orano/AREVA alone have absorbed €23.7 billion (US$202425.8 billion) or 86 percent of all public cash injections to industrial companies by the State’s Holdings Agency (APE) between 2011 and 2023.172

Energy Planning

The French government submitted a draft Multi-Year Energy Program (PPE) 2030–2035 for public consultation between 7 March and 5 April 2025.173 According to the government, there were 1,373 contributions to the debate.174

The plan does not count on any additional operating reactor by 2035 but, rather prudently, on a stable, annual production 360 TWh—EDF aims for 400 TWh by 2030—from the current fleet of 57 reactors. Operational lifetime extensions beyond 50 and 60 years and uprating possibilities would be assessed and implemented where possible and meeting safety criteria. No closures are foreseen in that period. The construction of “at least” a further 13 GW, equivalent to eight EPR2s, would be studied and the construction of “at least” one prototype SMR would be launched around 2030. And, just like in the 1970s, the plan envisions a “fleet of fast neutron reactors” operating with plutonium fuels “by the end of the century at the latest”.

However, the “Central Scenario” envisages strong growth for renewables, including more than doubling solar capacity to 54 GW generating 66 TWh by 2030 and expanding to 65–90 GW/92–110 TWh by 2035. Onshore wind is to grow by about half to 33 GW/72 TWh by 2030 and to 40–45 GW/91–103 TWh by 2035, with offshore wind to reach 3.6 GW/14 TWh by 2030 and 18 GW/71 TWh by 2035.

EDF is clearly opposed to high solar power targets and at the same time reveals that it is already experiencing (unspecified) constraints on the nuclear fleet never seen before due to variable generation demand:

The range mentioned in the consultation document on the PPE for solar PV, and even more so its upper limit of 100 GW175 in 2035, seems to us significantly too high for this timeframe. Such a level would lead to an imbalance in the French electricity production mix and in the supply/demand balance. In particular, the opportunities for nuclear power are already declining; thus, the sharp variations in power demand on the nuclear fleet over short periods are causing176 constraints on equipment and organizations that have never been encountered before.177

Conclusion

Nuclear production has improved in 2024 but remains far below the average level in the decade that ended in 2015. The fourth decennial inspections demand significant outage times, and new cases of stress corrosion cracking and thermal fatigue recreate uncertainty and additional inspection needs.

The overall number of outage days went down significantly in 2024, but the number of unplanned outage days (according to EDF definition) is the second highest level of the past six years.

The first EPR Flamanville-3 was finally connected to the grid in December 2024. The commissioning process has been difficult, and the unit has been off-grid most of the time due to excessive inspection needs, numbers of incidents, and repair work.

Consequentially, there is now no reactor under active construction in France.

The newbuild plans proceed, with bulldozers having started earthworks at the Penly site, while the Flamanville-3 EPR successor EPR2 still does not even exist on paper. That means that reliable cost estimates are hardly possible. EDF is to supply an updated calculation to the government by the end of 2025. The final investment decision is expected for the end of 2026. Construction start—first concrete for the foundation of the reactor building—has been delayed to September 2028.

The expected startup of the first EPR2 at Penly has been officially delayed from the originally planned 2035 to 2038.

Meanwhile, the renewables buildout is accelerating, especially with solar adding a record 5 GW in 2024. EDF is worried that the implementation of envisaged high solar capacity targets for 2030/2035 could threaten already “declining opportunities” for nuclear power and increase the need for reactors ramping up and down to follow fluctuating renewable output, which in turn has detrimental effects on equipment and organization.

The publication of the Multi-Year Energy Program (PPE) 2030–2035, supposed to frame energy policy for the next decade, has been delayed multiple times and is expected to happen once the National Assembly reconvenes after the summer break and gets a chance to comment.

Japan Focus

Overview

During Financial Year (FY) 2024, which runs from April 2024–March 2025, the number of nuclear reactors considered “operable” increased from 12 to 14 with a gross capacity of 13.3 GW.178 As no additional reactor has been declared for permanent closure during the past year, the total number of closed reactors remains unchanged at 27,179 including 22 units officially closed following the Fukushima accidents in 2011 (see Table 6 for details).

In November 2024, for the first time since the establishment of the new regulator in 2012, a nuclear reactor—Tsuruga-2—failed the restart safety review under the new regulations.

Total nuclear power generation increased by 9.5 percent, from 77.5 TWh in 2023 to 84.9 TWh (gross) in 2024. (See Figure 35) The share of nuclear power in the total power generation also increased from 7.7 percent in 2023 to 8.4 percent in 2024.180

1. Rise and Fall of the Japanese Nuclear Program

Sources: WNISR with Energy Institute and IAEA-PRIS, 2025

The current reactor fleet consists of 33 units (33.1 GW gross) of which 25 units (24.8 GW gross) have applied for an operating license under the new post-Fukushima regulations.181 Japan’s nuclear safety review process consists of three steps carried out simultaneously:182

1) Review of “changes in reactor installation” → Permission

Under the new regulatory standards, which incorporate lessons learned from the accidents at Tokyo Electric Power Company (TEPCO)’s Fukushima Daiichi Nuclear Power Station, licensing criteria have been strengthened for seismic events, tsunamis, and other hazards. These new standards are applied retroactively by backfitting existing reactors. Moreover, in cases where an accident or natural disaster exceeds the assumptions outlined in the revised criteria, countermeasures are required to prevent core damage, prevent containment vessel failure, and suppress the release of radioactive materials.

In the review for changes to the installation permit, the location, structure, and equipment of the nuclear reactor facility as well as the technical capabilities of the operator are evaluated for compliance with these standards.

2) Review of “plan for construction works” → Approval

This review verifies whether the detailed design of the nuclear reactor facility, as well as the methods of design and quality control during construction, are consistent with the installation permit and meet regulatory standards.

3) Review of “operational safety program” → Approval

This review assesses whether the safety measures stipulated in the reactor’s safety regulations are adequate for preventing disasters involving nuclear fuel materials, objects contaminated by nuclear fuel materials, and power reactors.

Generally speaking, passing the review of “changes in reactor installation” is regarded as passing the Nuclear Regulation Authority (NRA) review in Japan. However, the other two reviews must also be passed before the reactor can be put into operation.

So far, 17 units have passed the “changes in reactor installation” and the “plan for construction works” reviews while seven applications remain under review. The NRA is expected to issue an operating license for Tomari-3 of the Hokkaido Electric Power Company in the summer of 2025.

As of 1 July 2025, twelve of the 14 operable reactors were operating, while two—Ohi-3 and Takahama-4—were under periodic inspection since June 2025.183

WNISR considers 19 units in Long-Term Outage (LTO). In 2022, the IAEA adopted a new category called “Suspended Operation”. In 2024, it reclassified 19 reactors to be in Suspended Operation that WNISR considers in LTO. In other words, Japan and the IAEA have adopted an approach similar to the LTO concept that WNISR introduced in 2014. (See Figure 36).

1. Status of the Japanese Reactor Fleet

Sources: Various, compiled by WNISR, 2025

For the first time in 13 years since the Fukushima disaster was triggered, two Boiling Water Reactors (BWRs), Onagawa-2 and Shimane-2, resumed operation in late 2024.

As of mid-2025, the Japanese nuclear fleet consisting of 33 units, including 19 in LTO, had reached a mean age of 34.5 years, with 24 units aged over 31 years, of which five are over 41 years and one (Takahama-1) over 51 years old (see Figure 37).

1. Age Distribution of the Japanese Nuclear Fleet

Sources: WNISR with IAEA-PRIS, 2025

Regulatory Measures—First Rejection of Restart Application

The following overview summarizes events between mid-2024 and mid-2025.

Kashiwazaki Kariwa-6 and -7

On 27 February 2025, TEPCO announced a multi-year delay in targeted completion dates of the “Specific Major Accident Response Facilities”—designed in particular to counter terrorist acts—at Units 6 and 7 of the Kashiwazaki Kariwa nuclear power station (Niigata Prefecture) from September 2026 to September 2031 and from March 2025 to August 2029, respectively.184 Thus, the restart of Unit 7, which had been scheduled for the end of 2025, will be significantly delayed again.

Nuclear operators are obliged under the post-3/11 safety and security regulations to install these facilities at each nuclear power plant. The main facility has a backup function for a situation in which equipment would not be available in a wide area due to large-scale damage, such as following an intentional aircraft crash. There is a grace period of five years from the approval of the design and construction plan for the facility, with NRA-stipulated completion deadlines of September 2029 for Unit 6 and October 2025 for Unit 7. If the facilities cannot be installed within the grace period, the reactors will not be permitted to operate.

Units 6 and 7 passed the NRA’s review of “changes in reactor installation” in 2017, but in 2021 the NRA ordered TEPCO not to load fresh nuclear fuel at Kashiwazaki Kariwa after discovering deficiencies in anti-terrorism measures within the plant.185 The ban was lifted in 2023, and TEPCO has continued discussions with the Niigata Prefecture’s governor to obtain his consent to restart the plant. TEPCO cited a shortage of materials and staffing as factors contributing to the installation delay.186

In light of the current circumstances, Takeyuki Inagaki, the director of the Kashiwazaki Kariwa plant, held a press conference on 25 June 2025, announcing a change in strategy: the plant will no longer prioritize restarting Unit 7 and will instead focus on restarting Unit 6. Fuel loading for Unit 6 was completed on 21 June 2025. The plant still has a four-year grace period to install Specific Major Accident Response Facilities. During the press conference, Director Inagaki did not specify when Unit 6 would be restarted.187

Ohma

On 6 September 2024, J-Power, owner of the Ohma nuclear power plant in Aomori Prefecture, informed the town of Ohma that it will postpone the start of safety work, which was scheduled to begin in the second half of 2024.188

According to IAEA-PRIS, construction of the Ohma plant began in May 2010, as the world’s first commercial nuclear power plant designed to operate solely on MOX fuel, a mixture of plutonium and uranium extracted from spent nuclear fuel.189 However, construction work was halted after the 2011 events at TEPCO’s Fukushima Daiichi facilities.190

In October 2012, J-Power announced its decision to resume construction of the plant and stressed that it had “not yet set a date for the commencement of operation but intends to review the matter in the future based on progress in construction.”191 But construction activities were limited, until they were eventually stopped in 2021. The Ohma construction history to 2021 is presented in an IAEA report, which exposes that due to new regulations under discussion in 2012, “construction efforts were made within the limited scope that would not be affected by new regulations”, and states:

After the new regulatory standards became effective in July 2013, the progress of work has been practically restricted by factors such as the inability to have regulatory inspection (pre-service inspection).

At present [as of 2021], it was forced to preserve the equipment for a long period of time at the construction site and the vendors factories.

Since 2014, preservation measures have been implemented with consideration of long term preservation.192

According to a J-Power spokesperson, no construction work is being carried out as of mid-2025. The project is still under safety review by the NRA. A safety improvement plan to comply with the new regulatory standards was scheduled to be finalized this year but has been postponed to autumn 2026 when it is to be submitted to Aomori Prefecture and Ohma Town. Following that step, J-Power will seek final approval for the plan from the NRA, aiming to commence operation within fiscal year 2030.193

WNISR considers the reactor construction as suspended.

Onagawa-2

On 15 November 2024, Tohoku Electric Power Company’s Onagawa-2 (Miyagi Prefecture) was reconnected to the grid after several delays.194 It was the 13th nuclear reactor and the first BWR—the same basic design as the Fukushima Daiichi Nuclear Power Station’s reactors—to be restarted since the March 2011 Fukushima events. All other reactors in Japan that were previously restarted since 3/11 were PWRs. It is also the first time since the 2011 Great East Japan Earthquake that a nuclear power plant in the affected area has resumed operation.195 Work on safety upgrades at the Onagawa nuclear power plant began in May 2013. The 800-meter long seawall was raised to 29 meters above sea level as a tsunami response measure, and various measures have been taken in order to secure “electricity and cooling functions in case of severe accidents.”196 Tohoku Electric Power Co. applied to the NRA for a safety review in December 2013, and the review of “changes in reactor installation” was permitted in February 2020.197 Since then, in order to receive construction and operating permits, efforts were made to strengthen safety, including seismic reinforcement work. Implementation work was completed in May 2024. By late February 2023, Tohoku Electric Power Co. had passed the review of “operational safety program”. So it received the operating license.198 Tohoku Electric Power Co. must still install the “Specific Major Accident Response Facilities” by December 2026.199

Shimane-2

On 23 December 2024, Chugoku Electric Power Company reconnected Unit 2 at its Shimane plant, almost 13 years after it was shut down in 2012.200 About 450,000 people live within a 30-kilometer radius of the plant. Since the Fukushima events of 2011 began, Shimane-2 is the 14th reactor and only the second BWR to be restarted, following Tohoku Electric Power Co.’s Onagawa-2 in November 2024. In 2021, Shimane-2 passed NRA’s review of “changes in reactor installation”.201 Chugoku Electric Power Co. completed safety work in October 2024 and placed nuclear fuel in the reactor in November.202 Unit 2 is the only operating nuclear reactor in Japan located in a prefectural capital (Matsue-City).

Shimane-3

Chugoku Electric Power Company plans to start operation of Shimane-3 by the end of FY2030, and the unit is undergoing a safety review.203 On 3–4 April 2025, the NRA inspected Chugoku Electric Power Company’s Shimane-3. The inspection team checked the reactor containment vessel’s structure and its piping system which sends steam to the turbine. The NRA’s inspection of Shimane-3 is meant to provide a reference for future safety reviews.

At present, there is no active construction work at the unit, as it is undergoing NRA inspections under the new regulatory standards. Prior to the Fukushima Daiichi events in 2011, the containment vessel and the pressure vessel had already been installed within the reactor building. Fourteen years later, the primary operations are focused on quality-control inspections.204

WNISR considers the reactor construction as suspended.

Takahama-1 and -2

On 16 October 2024, the NRA approved the management policy for Kansai Electric Power Co. (KEPCO)’s Takahama-1 reactor (in the Fukui Prefecture) for the next 10 years.205 Unit 1 began operation in 1974. In May 2024, Unit 1, which began operation in 1974 and marked 50 years of operation on 27 March 2024, and Unit 2, which began operation on 17 January 1975 and has also passed the 50-year mark, were both approved to operate for up to 60 years.206

Tokai-2

On 23 August 2024, Japan Atomic Power Company (JAPC) announced a new postponement of the estimated completion of safety measures required for the restart of Tokai-2 in Tokai-mura, Ibaraki Prefecture, from September 2024 to December 2026, following the discovery of a deficiency in the foundation of the seawall under construction at the plant. At the same time, the company also informed the NRA, Ibaraki Prefecture, Tokai Village, and other surrounding municipalities about the discovery.207 JAPC indicated that it intends to do additional reinforcement work in the surrounding ground, leaving the foundation in place. The NRA still has to examine whether this additional work would meet the safety criteria. Tokai-2 passed the NRA’s safety review in 2018 for “changes in reactor installation” to meet regulatory standards set in 2012.208 But JAPC still needs construction and operating permits before restarting the reactor.

Tomari-3

On 30 April 2025, the NRA granted permission to review of “changes to the reactor installation” for Unit 3 at the Hokkaido Electric Power Company’s (Hokuden for short) Tomari nuclear power plant (Tomari-mura), which is effectively a certificate of acceptance for safety examination.209 Tomari-3 was shut down in May 2012. It is the newest nuclear reactor in Japan, having started operation in 2009 with a net design capacity of 866 MW. The draft of the review report went through a public comment process and was scheduled to be approved in the summer of 2025.210

In 2013, Hokuden applied to the NRA to restart Tomari-3, but it took the NRA more than 11 years to examine whether the faults on the site were active or not. One of the factors that prolonged the NRA review was the difficulty of proving the non-existence of active fault lines on the site. Reportedly, Hokuden was unable to provide convincing evidence primarily due to a lack of personnel with expertise. It took eight years to produce the evidence that there was no active fault with assistance from other electric power companies and experts. Hokuden is notably proceeding with the construction of a seawall to protect against tsunamis in order to receive the operating permit. It is aiming to restart operating Unit 3 in 2027.211

Tsuruga-2—Restart Application Rejected

On 13 November 2024, the NRA formally decided that the Japan Atomic Power Company (JAPC)’s Tsuruga-2 in Fukui Prefecture does not comply with regulatory safety standards. This is the first reactor that failed its restart safety review since the establishment of the NRA in 2012. Under the new regulatory safety standards developed after the Fukushima accidents were triggered in 2011, a reactor cannot be restarted if there is an active fault directly under the reactor. The NRA concluded that Tsuruga-2 did not meet safety standards because JAPC failed to prove that there was no active fault.212 At a press conference following the regular meeting, NRA’s Chairman Shinsuke Yamanaka emphasized, “We made a solid decision from a scientific and technical standpoint.” JAPC indicated its intention to submit a new application to the NRA in order to resume operations.213

Legal Cases Against the Restart of Reactors

Various legal cases against the operation of existing reactors continue. The following are three key decisions from the past year, all of which rejected the plaintiffs’ requests for injunctions.

Sendai-1 and -2 – Volcanic Eruption and Earthquake Risks

On 21 February 2025, the Kagoshima District Court ruled against the plaintiffs in a trial in which over 3,000 people, including local residents, demanded an injunction against the operation of Units 1 and 2 of the Kyushu Electric Power Company (Kyuden)’s Sendai nuclear power plant in Kagoshima Prefecture, on the grounds that “there is no concrete risk of an accident caused by volcanic eruption or earthquake.” In 2012, the plaintiffs had filed a lawsuit against Kyuden and the national government, demanding an injunction against the operation of the plant, claiming that “there are five calderas near the plant that have caused huge eruptions in the past, and the possibility of a catastrophic eruption cannot be ignored.” After a lengthy hearing, the Kagoshima District Court ruled that Kyuden’s assessment—that “the possibility of a catastrophic eruption occurring during the operational period of the plant is sufficiently small”—was based on a “reasonable scientific basis, consistent with the ‘Volcano Guide’ used by the NRA to assess eruption risk, and not unreasonable.” The plaintiffs reportedly intend to file an appeal with the High Court.214

Ikata-3 – Active Fault Line Below Site and Volcanic Risks in Question

On 5 March 2025, the Hiroshima District Court dismissed a lawsuit filed by local populations and others seeking an injunction against the continued operation of Unit 3 of the Shikoku Electric Power Company’s Ikata nuclear plant in Ehime Prefecture, on the grounds that no concrete danger of harm to the lives or health of the plaintiffs had been found. In addition, while the residents argued that there was a possibility of an active fault line near the plant but that an adequate survey had not been conducted, the court ruled that “there is no evidence to suggest that the survey conducted by Shikoku Electric Power Co. was inadequate.” Furthermore, the residents claimed that a pyroclastic flow could reach the vicinity of the Ikata facilities in case of an eruption of the same magnitude as the massive eruption that occurred 90,000 years ago at Mount Aso in Kumamoto Prefecture across the sea. In response, the Hiroshima District Court concluded, “It cannot be said that Shikoku Electric Power Co. underestimated the possibility of an eruption at Mount Aso.”215 The plaintiffs appealed the court decision.

Ikata-3 – Offshore Fault Line in Question

On 18 March 2025, a ruling was also handed down in another lawsuit aiming at the Ikata-3 restart license. The Matsuyama District Court ruled against the lawsuit, stating that there was no concrete danger to the plaintiffs’ lives or bodies. The plaintiffs had argued that Shikoku Electric Power Co. did not take into account the possibility of larger earthquake shocks stemming from the fault zone extending offshore to the north of the Ikata site. However, the court ruled that Shikoku Electric Power Co. made its evaluation based on its knowledge of lateral displacement faults and the results of further investigations, and no critical issues would have been identified.216

Spent Fuel Management

On 6 November 2024, Japan’s first offsite interim dry storage facility, for spent nuclear fuel generated by commercial nuclear power plants, began operations in Mutsu City, Aomori Prefecture. The facility is designed to store fuel from two electric power companies, TEPCO and JAPC, for up to 50 years. NRA issued an operating license to the facility’s operator, Recyclable Fuel Storage Company (RFS), the same day the inspection of the facility had been completed.217

In August 2024, RFS signed a safety agreement with Aomori Prefecture and Mutsu City.218 This agreement is not part of the NRA’s licensing process, but it is meant to build trust between the two parties. The agreement gives the governor and mayor substantial insight and monitoring possibilities, as well as veto power over project operation.

On 26 September 2024, 12 tons of fuel were delivered to this interim storage facility for the first time from TEPCO’s Kashiwazaki Kariwa plant. RFS plans to accept a total of 96 tons by the end of FY2026. Currently, a building with a maximum capacity of 3,000 tons has been completed, and another building will be added. In the future, the two buildings are designed to store up to 5,000 tons of fuel.219

Startup of Reprocessing Plant at Rokkasho Mura Delayed—Again

On 29 August 2024, Japan Nuclear Fuel Limited (JNFL), the operator of the spent nuclear fuel reprocessing plant under construction in Rokkasho Village, Aomori Prefecture, announced that it was postponing the plant’s target completion date from the end of September 2024 to the end of FY2026. The company also postponed the completion of its MOX fuel fabrication facility, which was also scheduled for completion by the end of September 2024, to the end of FY2027.220 The reprocessing plant is a facility for extracting plutonium from spent nuclear fuel for reuse in nuclear power plants and is at the core of the government’s official spent nuclear fuel management policy. It was planned to be completed by 1997, but due to repeated problems and tightened safety standards post-3/11, the completion date has been pushed back many times, with the latest delay reportedly marking the 27th postponement.221

KEPCO Planning to Ship Spent Fuel to France for Reprocessing

On 13 February 2025, Kansai Electric Power Company (KEPCO) submitted a new roadmap to Fukui Prefecture and the prefectural municipalities regarding the removal of spent nuclear fuel from three nuclear power plants operating in the prefecture. In October 2023, KEPCO had issued the Roadmap for Spent-fuel Measures to Fukui Prefecture and its plan to build a dry cask storage facility at three nuclear power plants. But because of the delay in the start of operations at JNFL’s reprocessing plant in Rokkasho, Aomori Prefecture, where the spent fuel is to be shipped for reprocessing, KEPCO needed to revise the roadmap. The new plan states that the amount of spent fuel to be shipped to France, to which the reprocessing of spent fuel will be outsourced starting in FY2027, would be doubled to 400 tons including 20 tons of MOX. The original plan announced in 2023 was to transfer 200 tons of spent fuel including 10 tons of MOX fuel to France.222

KEPCO operates seven reactors in Fukui Prefecture, the largest number in Japan. However, due to the repeated delays of the commissioning of the Rokkasho Mura reprocessing plant and the lack of dry storage facilities, spent fuel cannot be removed from the plants. Spent nuclear fuel continues to accumulate at the nuclear power plants and 87 percent of the storage capacity in pools had been saturated as of January 2025. Fukui Prefecture has insisted that KEPCO remove the spent fuel from the prefecture.223 The company anticipates that the reprocessing plant in Rokkasho Village, Aomori Prefecture, will begin operation by FY2027, and that a total of about 198 tons will be shipped there between the beginning of FY2028 and the end of FY2030.224

KEPCO emphasizes that if it follows the roadmap, the amount of spent fuel removed would eventually exceed the amount generated, and thus it estimates that the amount stored would decrease starting in 2033.225 But the utilization rate of the spent fuel pools of the seven reactors currently in operation is nevertheless expected to reach approximately 97 percent as of the end of FY2032 and stand still at 96 percent at the end of FY2034. Furthermore, the roadmap stipulates that the reprocessing plant must be completed and begin operation by the end of FY2026, and if the operation is delayed, KEPCO will be forced to revise its plans.

KEPCO had announced it would suspend the operation of Takahama-1 and -2, which have been in operation for more than 40 years, as well as Mihama-3, if Fukui Prefecture did not approve the plan, but on 24 March 2025, Fukui Prefecture announced it would accept KEPCO’s new plan.226

High Level Radioactive Waste Disposal Plans

The Nuclear Waste Management Organization of Japan (NUMO) conducts a three-stage survey to select a final disposal site to determine whether it is suitable for construction of a disposal facility: a “literature survey”, an “overview survey”, and a “detailed survey”. Of these, the literature survey analyzes documents and data, including ancient as well as updated documents, on the geology of the study area.227 Since high-level radioactive waste continues to emit strong radiation for a long period of time, the Japanese government intends to bury it deep underground for final disposal.

On 22 November 2024, NUMO released a report, constituting the first literature review in Japan on the selection of a final disposal site for high-level waste from reprocessing plants in geological formations, on the town of Suttsu and the village of Kamoenai, Hokkaido, was submitted to the mayors of these two municipalities. The report was also submitted to the Governor of Hokkaido.228

For the first time in Japan, the first phase in the selection of a disposal site, i.e. the literature review, had been in progress since 2020, covering the town of Suttsu and the village of Kamoenai.229

NUMO’s report concluded that the project could proceed to the overview study, which is the second step in the selection process for the final disposal site. In this study, sample gathering by bore holes is aimed at informing the geological survey, which requires the consent of the local government. But the Hokkaido Governor has reportedly expressed his opposition to the overview survey.230

In addition to these two municipalities, a literature review is being conducted since June 2024 in Genkai Town, Saga Prefecture, where the Kyushu Electric Power Co.’s Genkai plant is located.231 This is the first town with a nuclear power plant site that accepted the survey.

Closed Power Reactors in Japan

No additional reactors operating or in outage at the time of the Fukushima events, were formally declared for decommissioning in the year to 1 July 2025. The 22 Japanese reactors now confirmed to be closed and decommissioned total 15.5 GW or just under 35 percent of nuclear capacity prior to 3/11 (see Figure 35 and Table 6). In total, Japan has 27 closed reactors with 27.1 GW (see Japan in Decommissioning Status Report).

1. Official Reactor Closures Post-3/11 in Japan (as of 1 July 2025)

Sources: JAIF and JANSI, compiled by WNISR, 2025

Notes: This table only lists the 22 reactors closed after the Fukushima accidents, thus not including the Fugen Advanced Thermal Reactor (ATR), Japan Power Demonstration Reactor (JPDR), as well as Hamaoka-1 & -2 and Tokai-1.

BWR: Boiling Water Reactor; PWR: Pressurized Water Reactor; FBR: Fast Breeder Reactor; LTO: Long-Term Outage.

JAPC: Japan Atomic Power Company; JAEA: Japan Atomic Energy Commission

(a) – Unless otherwise specified, all announcement dates from JANSI, “Licensing Status for the Japanese Nuclear Facilities”, Japan Nuclear Safety Institute, 26 February 2020, see http://www.genanshin.jp/english/facility/map/, accessed 27 July 2020.

(b) – Unless otherwise specified, all closure dates from individual reactor’s page via JAIF, “NPPs in Japan”, Japan Atomic Industrial Forum,
see http://www.jaif.or.jp/en/npps-in-japan/, as of 27 July 2020.

(c) – Note that WNISR considers the age from first grid connection to last production day.

(d) – WNN, “Shikoku Decides to Retire Ikata 2”, 27 April 2018, see http://www.world-nuclear-news.org/C-Shikoku-decides-to-retire-Ikata-2-2703184.html, accessed 22 July 2018.

(e) – The Mainichi, “Japan Decides to Scrap Trouble-Plagued Monju Prototype Reactor”, 21 December 2016, see http://mainichi.jp/english/articles/20161221/p2g/00m/0dm/050000c, accessed 21 December 2016.

(f) – The Monju reactor was officially in Long-Term Shutdown or LTS (IAEA-Category Long Term Shutdown) since December 1995. Officially closed in 2017.

(g) – The decision to close the reactor was announced in October 2018.

New Energy Policy and the Role of Nuclear Energy

On 18 February 2025, the Japanese government approved the Seventh Strategic Energy Plan, the mid- to long-term energy policy guideline revised approximately every three years, during a cabinet meeting.232 In order to achieve carbon neutrality by 2050, the plan includes the objective to “maximize the use of renewable energy as our major power source” in the future, but also sets out to “maximize the use of both renewables and nuclear power.” Specifically, the plan sets the share of each power source in total power generation in 2040 at 40–50 percent for renewable energy and 20 percent for nuclear energy, which is not different from the previous target for 2030 as defined by the Sixth Strategic Energy Plan. Thermal power generation is still planned to represent 30–40 percent.233

The plan also claims the merits of nuclear energy: “Nuclear power has features such as excellent supply stability and technological self-sufficiency rate, cost levels comparable to other energy sources with little price fluctuation, and stable power generation at a constant output.”234 The goal to “reduce dependence on nuclear energy as much as possible”—explicitly stated in the Fourth, Fifth, and Sixth Strategic Energy Plans since the Fukushima events in 2011—has been deleted from the latest plan.235

The Seventh Strategic Energy Plan translates the shift in law and policy and clarifies the expected use of nuclear power generation. In 2023, Japan’s parliament passed the, so-called, “GX bundled bill”, which includes an amendment to the Nuclear Reactor Regulation Law, Electricity Utility Industry Law, and Atomic Energy Basic Law. Those three laws specify the main features of the new policy as follows:

Extension of the ‘licensing period’ (generally 40 years and 60 years for exceptional cases) allowing operators to apply for an extension corresponding to “certain shutdown period[s] due to ‘non-technical’ or ‘unplanned’ reasons.236

Thus, it can be said that the Seventh Strategic Energy Plan represents a fundamental change in energy policy in Japan since 2011. Independent experts, e.g., the Renewable Energy Institute, and environmental NGOs have opposed this new nuclear energy policy based in particular on the following arguments.

First, public opinion surveys (see next section) suggest that the majority of the population is still in favor of reducing dependency on nuclear power and eventual phaseout. Second, uncertainty regarding nuclear power’s expansion is too large to rely on. Thirdly, the target share of renewable energy is too low. The plan calls for a 40–50 percent renewable energy ratio by FY2040, which is much less ambitious than in other countries.237 Finally, economic competitiveness of nuclear power is not demonstrated and thus it would be less expensive to rely on renewable energy.

Even according to the Ministry of Economy, Trade and Industry (METI)’s own latest estimate, nuclear power is not necessarily the cheapest electricity source as suggested in previous estimates. The Agency for Natural Resources and Energy (ANRE) of METI now estimates that the cost per kWh in 2040, including government subsidies for each power source, would be ¥7.0–8.9 (US$0.05–0.06) for commercial solar, ¥7.8–10.7 (US$0.05–0.07) for residential solar, and ¥12.5 (US$0.09) for nuclear power. But ANRE suggests that renewable energy sources may need “adjustment for stabilization of power grid” (firming) which includes batteries and other technical equipment, and that if those costs were included, nuclear power would then have a cost advantage. According to the adjusted cost estimate, by 2040, the cost per kWh for nuclear power would be ¥16.4–18.9 (US$0.11–0.13)—unclear why nuclear goes up too—compared to ¥15.3–36.9 (US$0.11–0.26) for commercial solar and ¥19.5–25.2 (US$0.13–0.17) for onshore wind turbines.238 (For details on system issues, see Challenges of Integrating Nuclear Power into the Energy System).

Furthermore, the Seventh Strategic Energy Plan includes the promotion of the construction of “next-generation advanced reactors” with enhanced safety features. This is the first time since 3/11 that the Strategic Energy Plan explicitly mentions the construction of new reactors. As a specific measure, the plan allows electric companies that are close to decommissioning their nuclear reactors to build new next-generation innovative reactors on the sites of existing nuclear power plants they own.239

Prospects for Nuclear Power vs. Renewable Energy Deployment

The Seventh Strategic Energy Plan represents a major shift as it allows for the construction of new reactors in Japan for the first time since the Fukushima disaster began. It also aims to maximize the use of nuclear energy.

A recent public-opinion survey suggests split views on the use of nuclear energy. According to the survey published in March 2025 by Japan Atomic Energy Relations Organization, over 58 percent of respondents said they would accept the use of nuclear energy for the time being.240 However, there appears not to be much support for the construction of new nuclear reactors. In a public opinion poll conducted by public television station NHK in January 2025, only 21 percent of the respondents favored increasing the share of nuclear energy, a smaller fraction than those who prefer the “status quo” (30 percent) or who believe it “should be reduced” (31 percent).241 It remains less certain how the policy could impact the potential construction of new reactors. In addition, many issues associated with the decommissioning of the Fukushima Daiichi reactors remain unresolved (see Fukushima Status Report). Also, legal cases against restarting reactors and in favor of compensation for the impact of the Fukushima disaster continue. It remains unclear how Japan’s future nuclear energy policy will unfold.

According to Energy Institute, non-hydro renewables accounted 15 percent of total energy generation in 2024, (with solar at 9.5 percent, followed by geothermal at 4.5 percent, and wind at 1.1 percent). Including hydro, renewables accounted for 22.9 percent, with a small increase of 0.8 percentage point over the year. The production and share of solar almost tripled over the last decade, from 3.4 percent in 2015. Wind production and share more than doubled over the same period, but remained very low. The share of hydro has remained quite stable over the years at 7–8 percent.242 Japan remains heavily reliant on fossil fuels to generate electricity, as natural gas (31.3 percent), coal (29.6 percent) and oil (2.26 percent) combined covered 63 percent of power production in 2024.

Overseas, the introduction of wind power generation has made progress, especially in Europe, but it is slowing down due to rising construction costs, policy hurdles, and local opposition. A similar trend can be observed in Japan. Mitsubishi Corporation, a general trading company, has announced that it is reviewing its plans for offshore wind power generation off the coast of Akita and Chiba prefectures. Geothermal power generation also remains at a marginal 0.3 percent.243

To achieve the maximum electricity supply target from renewable energy sources for FY2040 in the Seventh Strategic Energy Plan, an increase of about 24 percent over the next 15 years would be required.

Russia Focus

Overview

In 1954, the Soviet Union was the first country to produce nuclear electricity for the grid. Since then, Russia has significantly influenced the global industry and turned—by far—into the largest exporter of reactors, currently constructing 23 large units244 (19 outside the country) and three small modular reactors of the world’s total of 63 projects in active building, as of 1 July 2025.

Russia has the fourth largest operating nuclear power fleet, with a total net capacity of 26.8 GW and 36 operational reactors across several designs: 10 pressure-tube Light Water Cooled Graphite Moderated Reactors (seven Chornobyl-type RBMKs and three EGP-6s), 24 PWRs (five VVER-440s, 13 VVER-1000s, four VVER-1200s, and two small KLT-40 “floating reactors”), and two Fast Breeder Reactors or FBRs (BN-600 and BN-800). Eleven reactors at different Russian power plants are closed: eight pressure-tube reactors (four RBMKs, two AMBs, one EGP-6, and one AM-1 reactor at Russia’s first plant in Obninsk) and three PWRs.

In 2024, nuclear energy contributed 17.8 percent of the country’s power mix, generating 216 TWh245 of gross electricity. This marks the second consecutive annual decline in gross nuclear electricity production from a record high of 223 TWh in 2022246 (see Figure 38). The share of nuclear power in Russia’s total electricity generation also declined for the fourth consecutive year from its peak of 20.3 percent in 2020, during the pandemic.247

1. Nuclear Production Versus Installed Capacity in Russia

Sources: WNISR with IAEA-PRIS and Energy Institute, 2025

Note: Capacity as of year-end.

The main reason for this decline is the country’s aging nuclear fleet with the gradual closure of end-of-lifetime units on one hand, and delays in constructing replacement capacities on the other hand. Since 2018, Russia has begun decommissioning the first Soviet-era gigawatt-class reactors. As shown on Figure 39, besides a small 10-MW reactor at Bilibino, four large first-generation RBMK units at the Leningrad and Kursk nuclear power plants—each with a service life of around 45 years—have already been closed. The last one was shuttered on 31 January 2024.248 But at these sites, only two new VVER-1200 units started up, at Leningrad II, to replace lost capacity. New units at Kursk II lag behind the closure of old reactors by at least four years. Several technical issues at new Russian units in 2023 also contributed to lower output and further reinforced the ongoing downward trend in production.249

1. Startups and Closures in Russia

Sources: WNISR with IAEA-PRIS, 2025

Regardless of the current stagnation in output, the State Atomic Energy Corporation Rosatom has also been given a task by Vladimir Putin personally—to achieve a 25 percent share of nuclear energy in the country’s electricity mix by 2045.250

Lifetime Extensions

As of mid-2025, the average age of Russian nuclear power units is 31.5 years (see Figure 40) which is only slightly below the global average. This intermediate age-position of Russia’s reactor fleet among the “Top 5 fleets”—between the younger fleets of China and South Korea, and the older ones of the United States and France (see Figure 17)—reflects Russia’s historical development. Part of its nuclear fleet is relatively old, with reactors built in the 1970s and 1980s, similar to those in the West. At the same time, construction activities in recent decades have helped lower the average age of Russian reactors (see Figure 18).

1. Age Distribution of the Russian Nuclear Fleet

Sources: WNISR with IAEA-PRIS, 2025

Nevertheless, 64 percent of all operating Russian reactors, including all RBMKs, all VVER-440s, and eight of 13 VVER-1000s, are running with extended lifetimes (beyond 30 years) (See Table 7). The oldest operating unit—Novovoronezh-4, a VVER-440— reached 53 years of service in 2025.

Due to technological challenges and economic impracticality, four first-generation RBMK reactors were denied lifetime extensions beyond 45 years and were closed upon reaching this limit between 2018 and 2024, and the remaining three EGP-6 units are expected to be closed by the end of 2025 after 50 years of operation. Meanwhile, several VVER units have already received licenses to operate for up to 60 years.

1. Status of the Russian Nuclear Reactor Fleet (as of 1 July 2025)

Reactor Model/ Reactor	Net Capacity (MW)	Grid Connection	Design Lifetime (in years)	Closure Date/ License Expiration Date	Age at Closure (End of License/ Planned Closure)(a)
LGWR
AM
APS-1 Obninsk	5	27/06/1954	n/a	Closed on 29/04/2002	47.8
Beloyarsk-1	102	26/04/1964	n/a	Closed on 01/01/1983(b)	18.7
Beloyarsk-2	146	29/12/1967	n/a	Closed on 01/01/1990(b)	22
EGP 6
Bilibino-1	11	12/01/1974	30	Closed in March 2018(c)	44.2
Bilibino-2	11	30/12/1974	30	Licensed to 12/2025(1)	(51)
Bilibino-3	11	22/12/1975	30	Licensed to December 2025(1)	(50)
Bilibino-4	11	27/12/1976	30	Licensed to December 2025(1)	(49)
RBMK(d)
Kursk-1	925	19/12/1976	30	Closed on 19/12/2021	45
Kursk-2	925	28/01/1979	30	Closed on 31/01/2024	45
Leningrad-1	925	21/12/1973	30	Closed on 22/12/2018	45
Leningrad-2	925	11/07/1975	30	Closed on 22/11/2020	45,4
Kursk-3	925	17/10/1983	30	Licensed to 31/12/2028(2)	(45/50)
Kursk-4	925	02/12/1985	30	Licensed to 21/12/2030(2,8)	(45/50)
Leningrad-3	925	07/12/1979	30	Licensed to 2030(3)	(50)
Leningrad-4	925	09/02/1981	30	Licensed to December 2025(4)	(45/50)
Smolensk-1	925	09/12/1982	30	Licensed to December 2027(5)	(45/50)
Smolensk-2	925	31/05/1985	30	Licensed to 2030(6)	(45/50)
Smolensk-3	925	17/01/1990	30	Licensed to December 2034(1)	(45/50)
PWR
VVER V-120
Novovoronezh-1	197	30/09/1964	20	Closed on 16 February 1988	24
Novovoronezh-2	336	27/12/1969	20	Closed on 29/08/1990	21
VVER V-440
Novovoronezh-3	385	27/12/1971	30	Closed on 25/12/2016	45
Kola-1	411	29/06/1973	30	Licensed to July 2033(1)	(60)
Kola-2	411	09/12/1974	30	Licensed to December 2034(1)	(60)
Kola-3	411	24/03/1981	30	Licensed to 2027(7)	(46)
Kola-4	411	11/10/1984	30	Licensed to December 2039(8)	(55)
Novovoronezh-4	385	28/12/1972	30	Licensed to December 2032(1)	(50)
VVER-1000
Balakovo-1	950	28/12/1985	30	Licensed to December 2045(8)	(60)
Balakovo-2	950	08/10/1987	30	Licensed to October 2043(1)	(56/60)
Balakovo-3	950	25/12/1988	30	Licensed to December 2048(1)	(60)
Balakovo-4	950	11/04/1993	30	Licensed to 2051(9)	(58/60)
Kalinin-1	950	09/05/1984	30	Licensed to June 2044(10)	(60)
Kalinin-2	950	03/12/1986	30	Licensed to November 2038(1)	(56/60)
Kalinin-3	950	16/12/2004	30	Licensed to October 2034(1,11)	(30/60)
Kalinin-4	950	24/11/2011	30	Licensed to October 2041(1,11)	(30/60)
Novovoronezh-5	950	31/05/1980	30	Licensed to 2035(7)	(55/60)
Rostov-1	989	30/03/2001	30	Licensed to January 2031(1,11)	(30/60)
Rostov-2	950	18/03/2010	30	Licensed to September 2039(1,11)	(29/60)
Rostov-3	950	27/12/2014	30	Licensed to November 2044(1,12)	(30/60)
Rostov-4	979	02/02/2018	30	Licensed to December 2057(1,11)	(39/60)
VVER-1200
Leningrad 2-1	1101	09/03/2018	60	Licensed to December 2057(1,11)	(29/60)
Leningrad 2-2	1101	23/10/2020	60	Licensed to July 2050(1,11)	(30/60)
Novovoronezh-2-1	1100	05/08/2016	60	Licensed to March 2046(1,11)	(30/60)
Novovoronezh-2-2	1101	01/05/2019	60	Licensed to February 2049(1,11)	(30/60)
KLT-40S “Floating”
Akademik Lomonosov-1	32	19/12/2019	40(13)	Licensed to June 2029(13)	(10/40)
Akademik Lomonosov-2	32	19/12/2019	40	Licensed to June 2029(13)	(10/40)
FBR
BN-600	560	08/04/1980	30(14)	Licensed to 2040(15)	(60)
BN-800	820	10/12/2015	40(14)	Licensed to 12/2043(1,16)	(28/60)

Sources: Various, compiled by WNISR, 2025

Notes:

(a) For operating reactors, age at the end of current valid license (as of 1 July 2025)/expected further extension.

(b) No or scarce production data available for Beloyarsk-1 and -2 in IAEA-PRIS; official closure from IAEA-PRIS but does not reflect the actual production of those reactors.

(c) Bilibino-1 was officially closed in January 2019, but according to IAEA-PRIS it was disconnected from the grid in March 2018; see IAEA-PRIS, “Operating Experience with Nuclear Power Stations in Member States – 2019 Edition”, see https://www.iaea.org/publications/13575/operating-experience-with-nuclear-power-stations-in-member-states.

(d) Planned extension to 50 years for the seven RBMKs.

Sources:

(1) Federal Environmental Industry and Nuclear Supervision Service and Rosatom, “The Ninth National Report of the Russian Federation on the Fulfillment of the Commitments Resulting From the Convention on Nuclear Safety To the Joint Eighth/Ninth Review Meeting of the Contracting Parties under the Convention on Nuclear Safety”, 2022, see https://www.iaea.org/sites/default/files/24/02/cns_8th_and_9th_rm_national_report_russian_federation.pdf.

(2) Rosenergoatom, “ОТЧЕТ по экологической безопасности за 2024 год”, Rosatom, 2025, see https://www.rosenergoatom.ru/upload/iblock/3ce/zimy1kcxqewvio5x2kcv6hcstlz3t2lb.pdf.

(3) Rosenergoatom, “Ленинградская АЭС получила лицензию Ростехнадзора на эксплуатацию энергоблока №3 до 2030 года”, Press Release, 3 February 2025, see https://www.rosenergoatom.ru/zhurnalistam/news/47889/, accessed 8 June 2025.

(4) Rosenergoatom, “Годовой отчет 2010”, 2011, see https://report.rosatom.ru/go/rosenergoatom/go_rosenergoatom_2010/go_rosenergoatom_2010.pdf.

(5) AK&M, “Smolensk NPP Has Received a License from Rostechnadzor for the Operation of Power Unit No. 1 until 2027”, 27 December 2022, see https://www.akm.ru/eng/news/smolensk-npp-has-received-a-license-from-rostechnadzor-for-the-operation-of-power-unit-no-1-until-20/; and Rosatom, “ОТЧЕТ по экологической безопасности Смоленской АЭС за 2023 год”, 2024, see https://www.report.rosatom.ru/go/2023/eco/rea/%D0%A1%D0%BC%D0%BE%D0%BB%D0%B5%D0%BD%D1%81%D0%BA%D0%B0%D1%8F_%D0%90%D0%AD%D0%A1.pdf.

(6) Rosenergoatom, “Смоленская АЭС получила лицензию Ростехнадзора на безопасную эксплуатацию энергоблока №2 ещё на 5 лет - до 2030 года”, Press Release, 2 June 2025, see https://www.rosenergoatom.ru/zhurnalistam/news/48487/.

(7) WNA, “Nuclear Power in Russia”, Updated 1 May 2025, see https://world-nuclear.org/information-library/country-profiles/countries-o-s/russia-nuclear-power.

(8) Federal Environmental Industry and Nuclear Supervision Service and Rosatom, “The Seventh National Report of the Russian Federation on the Fulfillment of Commitments Resulting From the Convention on Nuclear Safety”, 2016, see https://www.iaea.org/sites/default/files/russia-national-report-7th-rm-cns_en.pdf.

(9) Interfax, “Балаковская АЭС продлила на 28 лет лицензию на эксплуатацию энергоблока №4”, 22 December 2023,
see https://www.interfax.ru/russia/937618.

(10) Rosenergoatom, “Калининская АЭС получила лицензию Ростехнадзора на эксплуатацию энергоблока №1 в продленном сроке до 2044 года”, Press Release, 30 June 2025, see https://www.rosenergoatom.ru/zhurnalistam/news/48650/.

(11) See list of all Rosenergoatom licenses at Выставить-счет.рф, “Проверка контрагента—АО ‘КОНЦЕРН РОСЭНЕРГОАТОМ’”, 2025,
see https://выставить-счет.рф/vipiska-egrul/5087746119951/.

(12) Rosenergoatom, “Материалы обоснования лицензии на осуществление деятельности в области использования атомной энергии «Эксплуатация энергоблока №3 Ростовской АЭС в 18-месячном топливном цикле на мощности реакторной установки 104% от номинальной с вентиляторными градирнями”, see https://www.rosenergoatom.ru/stations_projects/sayt-rostovskoy-aes/bezopasnost-i-ekologiya/.

(13) Akademik Lomonosov Floating Nuclear Power Plant, “Floating Nuclear Power Plant Akademik Lomonosov Has Received an Operating License”, Rosatom, 28 June 2019, see https://fnpp.info/latest-news/floating-nuclear-power-plant-akademik-lomonosov-has-received-an-operating-license, accessed 15 August 2025.

(14) Ilia Pakhomov, “BN-600 and BN-800 Operating Experience”, Institute for Physics and Power Engineering, State Scientific Center of the Russian Federation, as presented at GEN IV International Forum, 19 December 2018, see https://www.gen-4.org/gif/upload/docs/application/pdf/2019-01/gifiv_webinar_pakhomov_19_dec_2018_final.pdf.

(15) Rosenergoatom, “Белоярская АЭС получила лицензию Ростехнадзора на продление эксплуатации энергоблока №3 до 2040 года”, Press Release, 1 April 2025, see https://www.rosenergoatom.ru/zhurnalistam/news/48170/.

(16) Rosenergoatom, “Environmental report of the Beloyarsk NPP for 2023”, 2024, see https://www.rosenergoatom.ru/upload/iblock/f36/xe0qu0bgew6ajj1umj678vhiu0uyg8p5.pdf.

Since late 2022, Rosatom has been considering new types of lifetime extensions not previously applied in Russia. One example is the planned five-year extension for seven second-generation RBMK reactors, bringing their operational life to 50 years. The first such license was issued in February 2025 for Leningrad-3.251

In 2025, a license was also issued to extend the operation of Russia’s oldest fast reactor, the BN-600, to 60 years.252 For the first time, an extension beyond 60 years has also been proposed for two VVER-440 units (now over 50 years old) at the Kola nuclear power plant,253 though no licenses have been granted yet.

The cost of current lifetime extension procedures is undisclosed, but past cases—like the RUB201217 billion (US$2012551 million) 15-year extension of Leningrad-4 in 2010254—suggest it is likely to reach hundreds of millions of dollars per unit.

Nuclear Newbuild

In its March 2021 strategic review, Rosatom stated that nuclear energy should supply 25 percent of the country’s electricity generation by 2045 in line with a target set by the President. According to Rosatom CEO Alexey Likhachev, this would require the commissioning of 24 units, including at new sites and in new regions.255

As of mid-2025, Russia has seven reactors under construction simultaneously: five nuclear power units, with capacities ranging from 300 MW to 1200 MW (two VVER-TOI units at Kursk II, two VVER-1200 units at Leningrad II, and the BREST-OD-300 SMR in Seversk), as well as at least one barge with two SMR modules. This is the highest number of reactors simultaneously under construction in Russia in a decade.

Two large units are under construction at Kursk II (Kursk 2-1 and 2-2) to replace old RBMK reactors, two of which have already closed. These VVER-TOI (VVER V-510) units with a net capacity of 1200 MW—the most powerful ever built in Russia—will be the first of the latest Russian designs (Generation III+), and they are also earmarked for export. When construction started on Unit 1 in April 2018 (see Annex 5), completion was scheduled for 2022 or 2023. Recent plans call for the unit to be commissioned by the end of 2025.256 Cost estimates for the two units have also reportedly almost doubled—from RUB225 billion in 2016257 to RUB459 billion in 2022 (US$20226.7 billion)258.

First concrete at Unit 3 of Kursk II is planned for December 2025.259 The corresponding license was issued in March 2025.260 Work on Unit 4 is expected to begin afterwards.

Another two large VVER-1200 units, Leningrad 2-3 and 2-4, are under construction at Leningrad II; these units are also called Leningrad-7 and -8 (when counting existing reactors at the site)261 and March 2025,262 respectively, with startups planned for Unit 7 in 2030 and Unit 8 in 2032.263 As in the Kursk case, new units at the Leningrad site are to replace aging RBMK reactors, with two already closed.

At the Smolensk nuclear power plant, which operates the three most recent RBMKs, site preparations for two VVER-TOI units began in 2024, with first concrete for the reactor building of Unit 1 planned for March 2027264 and commissioning of both units set for 2033 and 2035.265

Construction of an innovative small (300 MW) fast reactor design, the BREST-OD-300, using liquid lead as a coolant and uranium-plutonium nitride fuel started in June 2021.266 The reactor’s cost is said to be RUB2021100 billion (US$20211.4 billion). The objective for BREST-OD-300 is to start up in 2028, which, if achieved, would be impressive given that it is the first of its kind, but the project is not there yet.

Rosatom is involved in the construction of four small Floating Nuclear Power Plants (FNPP), each equipped with two RITM-200C reactors, to supply energy to mining projects in the Russian Arctic—specifically at Cape Nagloynyn in Chukotka. The first two “units” with two modules each are planned for commissioning by 2028, with the third and fourth to follow in 2029 and 2031, respectively.267

However, it is difficult to determine the exact number of floating nuclear units currently under active construction. Russia lacks established terminology and clear milestones for floating nuclear plants, such as “first concrete”—the beginning of concreting of the base slab of the reactor building—used for land-based reactors, making progress hard to track even in official reports and statements. Public information regarding progress on key components, such as reactor vessels and barge hulls, is also scarce, especially since at least one of the projects involves China, where information is also tightly controlled.

In 2021, Rosatom signed a US$226 million contract with China’s Wison Heavy Industry for two floating power plant hulls, with deliveries scheduled for October 2023 and February 2024.268 A keel-laying ceremony for the first barge was held in China in August 2022,269 and a third hull contract followed in 2024.270 However, no deliveries have been reported. In 2024, Wison’s parent company withdrew from Russian projects,271 likely due to concerns over secondary sanctions related to its involvement in the Russian Arctic LNG 2 project. It remains unclear how this decision will affect Rosatom’s contracts.

According to Russian News Agency TASS, Rosatom stated on 20 May 2025 that six RITM-200 reactors for FNPPs are concurrently being manufactured.272 This would suggest that three out of the four planned FNPPs are at some stage of construction. However, since the IAEA proposes to use the keel laying of the floating hull as the official starting point for the construction of a floating nuclear power plant, and given that reliable information is only available for the construction start of the first hull in August 2022, WNISR will consider only one floating power plant with two reactors as being under active construction.

At the end of December 2024, Rosatom’s specific proposals to achieve the goal to supply 25 percent of the country’s electricity generation by 2045 were formalized in the government-approved General Scheme for the Placement of Energy Facilities through 2042.273 According to the document, by 2042 Rosatom plans to have built 38 new nuclear reactors (see Table 8), including 12 SMRs, with a total capacity of around 294 GW. This is larger than the total capacity of Russia’s current nuclear fleet. According to the plan, by 2042 the share of nuclear energy in Russia’s electricity mix would be 24 percent.

1. Nuclear Reactors Under Construction and Planned in Russia

Plant	Reactor Design	Capacity Per Unit (MWe)	Reactors (Barges)	Capacity Per Site (MWe)	Construction Start	Planned Startup	Site
Kursk II	VVER-TOI	1 200	4	4 800	29/04/2018 15/04/2019	2025/27	Active
					No	2031/34
Leningrad II	VVER-1200	1 150	2	2 300	14/03/2024 20/03/2025	2030/32	Active
Novovoronezh II	VVER-optim	1 200	1	1 200	No	2036	Active
Beloyarsk	BN-1200M (fast reactor)	1 250	1	1 250	No	2034	Active
BREST ODEK	BREST-OD-300 (fast reactor)	300	1	300	08/06/2021	2028	New
Baimsky GOK (FNPP)	RITM-200C (2 per barge)	106	8 (4)	424	30/08/2022*	2028	New
					No	2029/31
Chukotka SMR	Shelf-M	10	1	10	No	2030	New
Kola II	VVER-600S	600	3	1 800	No	2035/37/40	Active
South	VVER-optim	1 200	2	2 400	No	2036/39	New
Primorsk	VVER-1000	1 000	2	2 000	No	2033/35	New
Norilsk SMR	RITM-400	80	4	420	No	2032/34/36/37	New
Yakutsk SMR	RITM-200N	55	2	110	No	2031	New
Smolensk II	VVER-optim	1 200	2	2 400	No	2033/37	New
Reft	RBN	1 255	1	1 255	No	2041	New
South-Ural	RBN	1 255	2	2 510	No	2038/40	New
Siberia	RBN	1 255	2	2 510	No	2041/42	New
Seversk	RBN	1 255	2	2 510	No	2037/39	New
Khabarovsk	VVER-600S	600	2	1 200	No	2041/42	New
Under Construction			7 (1)	5 106
Planned			35 (3)	24 293

Sources: General Scheme for the Placement of Energy Facilities through 2042, with Rosatom and IAEA-PRIS, 2025

Note: *Keel laying of one barge designed to host two reactors, counted as two reactors under construction.

The document notes that several SMR projects, such as those in Chukotka, Norilsk, the second unit in Yakutsk and even the Baimsky GOK, will only be confirmed after power purchase agreements are signed with end users. In other words, these projects may not move forward, or may proceed only partially, if such agreements are not secured.

Nearly half of the new capacity is planned over the next decade at existing sites in western Russia, with new sites in central and eastern regions expected from the 2030s. The focus remains on large VVER reactors (18 units), while the fast reactor fleet is projected to grow from two to about nine units by 2042. Most SMR projects are based on RITM reactors, already used on eight nuclear icebreakers.

Rosatom primarily builds new large nuclear units under the Capacity Supply Agreement (CSA) mechanism. After commissioning, the unit receives increased capacity payments on the wholesale market to ensure accelerated return on investment. For earlier nuclear projects, contracts allowed for cost recovery over 20 years.274

In 2023, according to the newspaper Kommersant, Russia’s Ministry of Energy began optimizing the capital and operating costs of new nuclear power plants by proposing to extend the duration of existing contracts to 25 years, with a baseline return rate of 10.5 percent. At the same time, it was proposed to limit CAPEX at around RUB184,100/kW (~US$20212,500/kW), which is reportedly in line with the cost of previous VVER-1200 units, in particular those put into operation at Novovoronezh II in 2016 and 2019,275 and would be close to the latest estimate of the projected cost of the nearly completed Kursk 2-1—RUB182,900/kW (US$20222,670/kW).276 At the same time, the single-rate price for the Kursk II nuclear power plant under the CSA agreement may amount to around RUB7/kWh277—equivalent to approximately US$75–102 per MWh, depending on exchange rate fluctuations in recent years.

Reactor Exports

Rosatom is the biggest nuclear power technology provider in the world; it is constructing 27 out of the 63 units under construction worldwide as of 1 July 2025. Of these 27 units, 20 are being built abroad in seven countries—including Mochovce-4, a Russian design completed by a Czech-led consortium in Slovakia—and seven at home, including SMRs. As of mid-2025, Rosatom or one of its subsidiaries is involved as the main contractor or vendor in the following projects abroad which are in various stages of active construction (see Annex 5 for further references):

• Rooppur, Bangladesh. Construction started on two VVER-1200 reactors at Rooppur in 2017 and 2018, which were expected to begin operation in 2023 and 2024, respectively.278 The first fuel load shipment was delivered in October 2023,279 but the first unit was not launched in either 2023 or 2024. In March 2025, the head of Rosatom stated that commissioning the first unit of the Rooppur nuclear power plant is one of the corporation’s key priorities for the current year.280 The project is facing challenges in particular due to underdeveloped local grid infrastructure and accumulated debts resulting from U.S. sanctions on Russia.281 See section on Bangladesh.
• Tianwan and Xudapu, China. Two VVER-1200s each at Tianwan and Xudapu (or Xudabao) are being built. Construction of the respective first units (Tianwan-7 and Xudapu-3) started in 2021 and that of the second pair (Tianwan-8 and Xudapu-4) in 2022.282 The commissioning of the units is scheduled for 2026–2027 (Tianwan)283 and 2026–2028 (Xudapu).284 China’s projects are the most localized among Rosatom’s constructions abroad; the company only designs and supplies core reactor island equipment, while Chinese partners handle all other aspects including project management.285 See China Focus.
• El Dabaa, Egypt. Four VVER-1200s are under construction at El Dabaa. This is Rosatom’s most recent construction project—all four units began construction from July 2022 onwards, with the latest starting in January 2024. The construction of the four reactors is scheduled to be completed by 2030. The cost is given as US$30 billion, facilitated by a US$25 billion loan from Russia.286 See section on Egypt.
• Kudankulam, India. Four VVER-1000s (Units 3 to 6) are under construction at Kudankulam. Construction started on the first pair in 2017 and the second pair in 2021. According to the site director’s estimates as of March 2025, the commissioning of the first of these units is expected in 2026. In December 2023, Russia and India signed “two important amendments, which will take the Kudankulam nuclear power project forward,”287 but the details are not disclosed, and the meaning remains unclear. According to the roadmap for nuclear cooperation between Russia and India, signed over a decade ago, a total of up to 12 reactors were envisaged to be built in India, including up to six in Kudankulam.288 See section on India.
• Bushehr, Iran. Construction of Bushehr nuclear power plant with two units was initially started in 1976 by the German company KWU-Siemens and was suspended in 1979. In 1995, the Russian company Atomstroyexport was contracted to complete and reconstruct the first unit of Bushehr, equipped with a VVER-1000 reactor. The reactor was connected to the grid in 2011. In 2014, a new US$10 billion contract was signed for the construction of two additional VVER-1000 units.289 New basemat concrete for the second unit (Bushehr-2) was poured in 2019, and its commercial operation is expected in 2029. Concrete pouring for the third unit was expected in 2024,290 but no official updates have been reported. Russia and Iran are also discussing the possibility of building additional units.291 The roadmap signed in 2014 allows for the construction of two more units at the Bushehr site and an additional four units at a new site.292 See section on Iran.
• Akkuyu, Türkiye. Four VVER-1200s are being built at Akkuyu. Construction started on them between 2018 and 2022. This is Rosatom’s first project being built under a BOO (Build-Own-Operate) scheme with the plant owned by Rosatom. The first unit was originally scheduled for commissioning in 2023, but due to a range of technical, political, and other reasons, including U.S. financial sanctions and delays in the delivery of Western equipment, in particular from Siemens, after 2022, commissioning has been repeatedly postponed. Unit 1 is currently expected to be commissioned in 2026.293 See section on Türkiye.

In addition, preparatory work is underway at Hungary’s Paks II plant for two VVER-1200s. The European Commission approved the contract in May 2023 despite the ongoing conflict between the E.U. and Russia on energy and the war in Ukraine.294 On 24 June 2024, the E.U. adopted its 14th sanctions package against Russia, which included a full exemption for Hungary’s Paks II nuclear project under Article 12h.295 Back in March 2021, prior to the full-scale invasion of Ukraine, Rosatom had planned to pour first concrete at the Paks II site in 2022.296 However, as of mid-2025, this has still not happened. Rosatom has stated it was “doing everything” to officially start construction before the end of 2025.297 The situation surrounding the project remains rather uncertain, as there are external factors both hindering its progress, e.g., attempts to challenge the project’s legality within the E.U.,298 and supporting its advancement, e.g., the recent easing of U.S. sanctions.299

Rosatom is a subcontractor in Slovakia’s Mochovce-4 project—a Soviet-era VVER unit whose construction began in 1985 and is now set for completion in 2025 by an international consortium. Rosatom provided technical support, while the main equipment came from Czech company Škoda JS. The twin Mochovce-3, connected to the grid in 2023, became the E.U.’s 19th VVER unit based on Russian design.

Most of these projects are long-standing, with contracts signed over seven years ago, and Rosatom has not secured any new foreign nuclear power plant construction contracts in recent years. The only exception is a May 2024 contract with the State Unitary Enterprise “Directorate for the Construction of Nuclear Power Plants” under Uzatom, the Atomic Energy Agency Under the Cabinet of Ministers of the Republic of Uzbekistan, to build six 55-MW SMRs in Uzbekistan,300 where a larger VVER-1200 project had also been under negotiation. While the shift to SMRs was not Rosatom’s original goal, the larger project remains under discussion. Still, the deal marks Rosatom’s first international SMR contract—a significant milestone.301

In June 2025, Rosatom was also selected to lead an international consortium for the construction of Kazakhstan’s first nuclear power plant.302 However, the contract has not yet been signed, and it remains unclear which other companies will join the consortium and how roles will be distributed among the partners. Meanwhile, it has been decided that the construction of Kazakhstan’s second nuclear power plant would be entrusted to China.303

Nuclear Interdependencies and Sanctions

Since its full-scale invasion of Ukraine in February 2022, and as of summer 2024, Russia was facing sanctions from forty-five countries—mainly in Europe and North America, but also from countries like Australia, Japan, New Zealand, and South Korea.304 As of mid-2025, further sanctions were under discussion. Ukraine is likely the only country that has sanctioned more than two hundred Rosatom-linked companies (from more than 400) and cut all ties with the corporation. Western countries have taken a more selective approach, targeting mostly military-linked entities and companies directly connected to Russia’s occupation of the Zaporizhzhia nuclear power plant, for example, the joint stock company “Operating organization of Zaporizhzhya NPP”, established by Rosenergoatom to manage the occupied nuclear power plant.305

By early 2025, nearly 70 Rosatom-related subsidiaries and individuals were under U.S. sanctions,306 with some also sanctioned by the E.U., U.K., Canada, and others. However, other than the termination of the contract for the Finnish Hanhikivi-1 project, so far, these sanctions have not jeopardized Rosatom’s pre-2022 foreign newbuild contracts, even if they have caused some difficulties and delays. Rosatom’s strong global presence, the challenge of replacing one of only a handful of nuclear suppliers, as well as dependencies on Russia as a client for Western equipment, appear to have given Rosatom a kind of sanctions immunity. See dedicated chapter on Russia Nuclear Interdependencies for details.

The European Union

Since 2022, and as of 1 July 2025, the E.U. has adopted 17 sanctions packages against Russia, targeting finance, energy, technology, and trade.307 Measures include disconnecting banks from SWIFT, freezing Central Bank assets, banning imports of oil, coal, gold, and diamonds, and restricting high-tech and dual-use exports. However, despite these restrictions, the E.U. continued to import unsanctioned fossil fuels, such as pipeline gas and LNG, from Russia, amounting to €21.9 billion (US$23.7 billion) in 2024 according to some estimates, with the volume decreasing by only 1 percent compared to the previous year.308 Half of this amount was accounted for by the five largest buyers—Hungary, Slovakia, Italy, France, and Czechia.309

WNISR2024 contained a brief overview of how nuclear equipment and fuel remained largely exempt from E.U. sanctions due to concerns and political pressure from certain Member States—notably France and Hungary310—with strong nuclear ties to Russia. Hungary continues work on Paks II, while France appears committed to its nuclear cooperation with Russia, including enriched uranium imports, joint projects with Rosatom like fuel fabrication in Germany, and sales of its Arabelle turbines for VVER reactors.

On 6 May 2025, the European Commission presented its roadmap for reducing reliance on Russian energy supplies, including nuclear supplies.311 The development of a step-by-step phaseout plan for Russian nuclear imports in cooperation with Member States by the end of 2025 revolves around proposed measures which include:

• economic incentives and tariffs to make the purchase of Russian enriched uranium less attractive;
• a ban on prolongations and new contracts with Russian suppliers (the Euratom Supply Agency that must counter-sign any such contract will not do so anymore).

The introduction of this roadmap—covering issues omitted from the E.U.’s 17 previous sanctions packages—is a positive step. However, much depends on the still-unspecified details. There are concerns about continued cooperation with Rosatom through loopholes such as licensed fuel production in the E.U. (e.g., Framatome’s Lingen facility, see Framatome and the Lingen VVER Fuel Manufacturing Plant Project312). It also remains unclear how the measures will be adopted amid opposition from Member States like Hungary and France.

One of the E.U.’s main dependencies on Russia in the nuclear sector is the use of Russian nuclear fuel for 19 VVER reactors operating in five countries: the Czech Republic, Hungary, Slovakia, Bulgaria, and Finland. As shown in WNISR2024, according to the Norwegian organization Bellona’s analysis based on Eurostat data, purchases of such fuel nearly doubled in the first year (2023) after the start of the all-out war with Ukraine, as operators sought to stockpile supplies amid increasing uncertainty. However, these purchases significantly declined from 635 tons (€718 million or US$2023776 million) in 2023 to 438 tons (€527 million or US$2024572.8 million) in 2024, although this still exceeds the pre-war level of around 300 tons per year.313 The 2023 total of 635 tons includes 62.3 tons supplied to Bulgaria, according to Eurostat, but data on Bulgaria’s imports in other years is missing, suggesting the database may be incomplete.

By 2024, all five countries using Russian nuclear fuel had signed contracts with alternative suppliers (Westinghouse and/or Framatome), and in Bulgaria and Finland, non-Russian fuel was already being loaded into reactors that year.314 (See Russia Nuclear Interdependencies). On 30 May 2025, the first batch of Westinghouse fuel was also delivered to Temelín in the Czech Republic.315

France remains virtually the only buyer in the E.U. of Russian enriched uranium

Imports of enriched uranium from Russia likewise fell from 253 tons (€430 million316 or US$2023465 million) in 2023 to 147 tons (€170 million or US$2024185 million) in 2024, with France remaining virtually the only buyer.317 That material was delivered either directly to France or to Framatome’s fuel manufacturing facility in Lingen, Germany.

The United States

For several decades, the U.S. has remained one of the main buyers of Russian enriched uranium, sourcing up to a third of its needs from Russia.318 In 2023, amid ongoing uncertainty, purchases even reached their highest level since 2013—over 700 tons worth US$1.2 billion.319

A new law banning the import of uranium from Russia was passed in May 2024 and came into effect on 11 August 2024.320 Given the significance of Russian supply, a caveat allows for waivers until the end of 2027, if the U.S. Department of Energy judges that no alternative supply is available. The legislation also unlocks $US2.7 billion in federal support for domestic enrichment.321

In response, in mid-November 2024, the Russian government imposed a temporary ban on enriched uranium exports to the U.S. until the end of 2025 and revoked existing export licenses. However, like the U.S. ban, this restriction can be bypassed through the issuance of special one-time licenses. As a result, enriched uranium deliveries from Russia to the U.S. resumed in early 2025.322

Nevertheless, by the end of 2024, exports of enriched uranium from Russia to the U.S. had nearly halved compared to the previous year—down to 335 tons worth US$624 million.323 Judging by a series of measures and executive orders issued by the Trump administration in May 2025 aimed at supporting the U.S. nuclear industry, the current administration appears set to continue efforts to reduce dependence on Russian supplies.324 For further details on U.S. policy, see United States Focus.

South Korea Focus

Republic of Korea (hereafter, South Korea) became a nuclear-operating country in 1977, when its first commercial reactor, Kori-1, was connected to the grid. The first reactor was a Westinghouse-designed Pressurized Water Reactor (PWR, WH 60), marking the country’s dependence on foreign technology during the initial phase.

During the 1980s and 1990s, South Korea pursued an aggressive technology transfer (import) strategy, culminating in the development of its own standardized reactor designs. The OPR-1000 was developed drawing on the System 80 design of Combustion Engineering (C-E), a U.S.-based nuclear engineering company, through a formal technology transfer agreement.325 C-E was acquired by Westinghouse in 2000, but the collaboration laid the foundation for Korea’s reactor localization. This evolution continued with the APR-1400, which further built upon the System 80+ framework and integrated significant domestic design innovations.

By 2007, South Korea surpassed Germany in nuclear electricity generation, rising from sixth to fifth place globally. Following the Fukushima events in 2011, Japan significantly curtailed its nuclear operations, and in 2012, South Korea temporarily became the fourth-largest nuclear power producer in the world. However, this position was overtaken by China in 2015, as the country rapidly expanded its nuclear fleet. Since then, South Korea has remained in fifth place globally in terms of both operating capacity and annual nuclear electricity generation. As of 1 July 2025, South Korea’s installed nuclear capacity exceeds that of the 15 smallest nuclear fleets in the world combined.

To date, two reactors have been closed: Kori-1 in 2017 and Wolsong-1 in 2019. The two oldest reactors in the country (Kori-2 and Kori-3) were shut down in April 2023 and September 2024, respectively, after their operating licenses expired. As they are expected to restart, they are considered in LTO, leaving 24 reactors in operation. In addition to Saeul-3 and -4 (formerly known as Shin-Kori-5 and -6), under construction since 2017 and 2018, respectively, construction started in May 2025 at Shin Hanul-3. This places South Korea among a small number of countries that are not only maintaining but also expanding their nuclear fleet.

Among the 31 countries with operating reactors, South Korea stands out for its highly standardized fleet, high load factors, and centralized management through Korea Hydro & Nuclear Power (KHNP), a subsidiary of state-owned Korea Electric Power Corporation (KEPCO).

In 2009, South Korea became a nuclear-exporting country by securing a contract to construct and operate four APR-1400 reactors at the Barakah site in the United Arab Emirates (UAE). While the Barakah project in the UAE remained its only full-scale export deal for over a decade, on 4 June 2025 South Korea finalized a contract to supply two APR-1000 reactors for the Dukovany expansion project in the Czech Republic. If successfully completed, this would mark South Korea’s first entry into the European nuclear market.

Policy Overview: From Nuclear Expansion to Political Uncertainty

From the second half of 2024 to the first half of 2025, South Korea underwent a dramatic political upheaval with significant implications for its energy policy landscape. Since 2022, under the Yoon Suk-yeol administration, the country had enhanced pro-nuclear policies, integrating nuclear power into national energy, industrial, and export strategies. However, on 3 December 2024 then-President Yoon declared an unconstitutional and unlawful emergency martial law, sparking widespread public backlash and cross-party condemnation. The National Assembly impeached him on 14 December 2024, and the Constitutional Court upheld the impeachment on 4 April 2025, permanently removing Yoon from office. This abrupt change created a sudden rupture in nuclear-policy leadership, introducing considerable uncertainty about the future direction of South Korea’s nuclear program.

Yoon’s rise to power was significantly aided by his vocal criticism and political attacks on the nuclear phaseout policy pursued by former President Moon Jae-in, which played a major role in mobilizing conservative support and shaping Yoon’s pro-nuclear platform.326 Once elected, Yoon strongly reversed course on the nuclear phaseout, prioritizing nuclear energy expansion instead.

A key expression of President Yoon’s pro-nuclear agenda was the 11th Basic Plan for Long-term Electricity Supply and Demand (BPE), which was formally established by the Ministry of Trade, Industry and Energy (MOTIE) on 21 February 2025 and later revised and re-announced on 13 March 2025.327 The 11th BPE proposed a long-term electricity mix in which nuclear power would supply about 32 percent of electricity by 2030 and just over 35 percent by 2038, up from 31.7 percent in 2024 (see Table 9). This includes plans for two new 1400-MW (APR-1400) reactors and four SMRs totaling 0.7 GW (four 170 MW i-SMRs328), although the i-SMR still remains at the Research and Development (R&D) phase and has not received standard design approval yet.329

1. Power Generation by Source and Share Forecast

	Unit	Nuclear	Coal	LNG	Renewable Energy	New Energy	Clean Hydrogen & Ammonia	Other	Total
Actual Electricity Mix in 2024	TWh	188.8	167.2	167.2	53.3	9.5	-	9.6	595.6
	Share	31.7%	28.1%	28.1%	8.9%	1.6%	-	1.6%	100%
11th Basic Plan on Electricity Supply and Demand - Electricity Mix Targets
2030 Target	TWh	204.2	110.5	161.0	120.9	18.7	15.5	11.8	642.6
	Share	31.8%	17.2%	25.1%	18.8%	2.9%	2.4%	1.8%	100%
2038 Target	TWh	248.3	70.9	74.3	205.7	26.4	43.9	34.9	704.5
	Share	35.2%	10.1%	10.6%	29.2%	3.8%	6.2%	5.0%	100%

Sources: KEPCO and MOTIE, 2025330

In addition to newbuilds, the Yoon administration institutionalized long-term reactor operation. Amendments to the Enforcement Decree of the Nuclear Safety Act (effective from late 2022) allowed operators to apply for lifetime extensions five to ten years before license expiry, streamlining the process and aligning it with industry planning cycles.331 The nuclear industry presumed all existing reactors would be extended,332 provided the independent Nuclear Safety and Security Commission (NSSC) grants permission.

However, with the impeachment and removal of President Yoon, the direction of nuclear policy has entered a new phase of uncertainty. The incoming President Lee Jae-myung from the Democratic Party of Korea, who took office in early June 2025, has not yet provided detailed clarification on how he will approach nuclear policymaking.

During a pre-electoral televised debate, Lee stated:

While we cannot completely eliminate nuclear energy overnight, the era of building more reactors should come to an end. (…)

We need to act urgently to strengthen our renewable base—not just for the environment, but for our economy to survive global RE100 [100% Renewable Energy] standards. (…)

Nuclear energy may appear cheap, but when you account for what happens if something goes wrong, it becomes one of the most expensive options.333

These and similar remarks suggest a cautious, pragmatic stance—neither fully endorsing nor rejecting nuclear energy outright. In contrast, the conservative candidate Kim Moon-soo from the People Power Party, who placed second in the presidential election, had pledged to increase nuclear’s share to 60 percent334—35 percent from large reactors and 25 percent from SMRs. However, his final published campaign manifesto lowered the target to 35 percent.335

It is difficult to understand what such percentage-based nuclear targets mean in terms of actual reactor capacity. To achieve the 35-percent nuclear share by 2050 proposed in Kim Moon-soo’s revised manifesto, South Korea would need to construct an additional 24 large-scale 1400-MW reactors336—on top of all currently operating, under construction, and planned units—even if the lifespan of each reactor were extended twice. This appears to represent an impossible target from a purely industrial point of view, considering past and present experience.

As of 1 July 2025, it remains unclear whether the 11th BPE could and would be implemented as outlined. The newly launched Lee Jae-myung administration has not announced any specific vision or clear plans regarding policy directions for power generation, including nuclear power. President Lee promised during his presidential campaign to merge the Ministry of Trade, Industry and Energy’s department of energy affairs with the Ministry of Environment’s department of climate crisis response into a new integrated “Ministry of Climate and Energy.”337 As such, it is expected that the incoming government’s power policy will take shape during the process of establishing the new ministry and appointing its minister.

Operating Fleet

As of 1 July 2025, South Korea operates 26 commercial nuclear reactors at five sites: Kori, Saeul, Hanul, Wolsong, and Hanbit. Including two reactors currently in LTO, the total operating capacity stands at approximately 24 GW. The reactor fleet is highly standardized, with 23 of the 26 units based on variations of the Pressurized Water Reactor (PWR) technology, while only three units—Wolsong-2, -3, and -4—use the Canadian CANDU pressurized heavy water reactor (PHWR) technology.

Reactors Under Construction and Planned

Three reactors—Saeul-3 and Saeul-4 as well as Shin-Hanul-3—remain under construction as of 1 July 2025.

The two Saeul APR-1400 reactors, located in Ulsan, were first included in the 4th BPE, established in December 2008. At that time, the reactors were scheduled for completion by December 2018 (Saeul-3) and December 2019 (Saeul-4).338 The construction licenses for both units were granted on 27 June 2016. First concrete was poured for Saeul-3 in April 2017 and for Saeul-4 in September 2018.339 As of the most recent 11th BPE, the expected completion dates for the two reactors are now set for February 2026 (Saeul-3) and November 2026 (Saeul-4).340

In addition to Saeul-3 and Saeul-4, two more reactors have received construction permits: Shin-Hanul-3 and Shin-Hanul-4, both also based on the APR-1400 design. The NSSC granted construction licenses for both units on 12 September 2024. According to a press release from KHNP, construction on Shin-Hanul-3 has officially begun on 20 May 2025 with the first concrete pour for foundations of the reactor building.341 KHNP President Whang Joo-ho stated that “we will do our utmost to ensure the safe construction of Shin-Hanul Units 3 and 4 and to achieve the goal of delivering the project on time and within budget, thereby enhancing the reputation of K-Nuclear in the global market.”342

Given that all eight reactors—Shin-Kori-2, Shin-Wolsong-1, Shin-Wolsong-2, Saeul-1, Saeul-2, Shin-Hanul-1, and Shin-Hanul-2—that were connected to the grid in South Korea since 2012, along with the four Barakah units in the UAE, were increasingly delayed from their original timelines, the promise of being “on time” appears very optimistic.

The construction of two additional large-scale APR-1400 reactors was introduced in the 11th BPE, scheduled for completion in June 2037 and June 2038 respectively, though no sites have yet been designated for them. The same plan also outlines the expected completion of four 175-MW SMRs in September and December 2034 and in March and June 2035. If all of these plans are realized and the lifespans of existing reactors are extended, the number of operating reactors in South Korea would increase from 26 (with the restart of Kori-2 and Kori-3) to 36, and the total installed nuclear capacity would expand from approximately 26 GW to 35 GW by 2038.

Small Modular Reactors (SMRs)

SMRs have emerged in recent years as a topic of significant public interest in South Korea. According to a search on Big Kinds,343 a news database provided by the Korea Press Foundation, covering over 100 national and regional news outlets of various kinds, the number of news articles mentioning SMRs increased steadily from 5,576 in 2022 to 6,408 in 2023, reaching 7,717 in 2024. Media coverage has primarily focused on investment and industrial strategy aspects.

Media focus on SMRs is partly due to the participation of KHNP in the government-led innovative SMR (i-SMR) development project, jointly promoted by the Ministry of Trade, Industry and Energy and the Ministry of Science and ICT (MSIT), with the Korea Atomic Energy Research Institute (KAERI) as a partner. The i-SMR is currently under development with the goal of obtaining standard design approval in 2028.344 Reportedly, KHNP has conducted preliminary preparations in the first half of this year to select a construction site for the i-SMR and plans to launch a site-solicitation process in the second half.345

In addition, major South Korean conglomerates are involved in U.S.-based projects—such as Doosan Enerbility, Samsung C&T Corporation, and GS Energy in NuScale or SK Group and HD Hyundai with TerraPower—as investors, equipment and service providers, or future strategic partners.346 KHNP is also reported to be investing in SMRs by purchasing shares in the special purpose company (SPC) established by SK Group to acquire the latter’s stake in TerraPower.347 According to the minutes of KHNP’s 5th Board of Directors meeting held on 28 August 2024, the draft Plan for Acquiring Shares in a U.S. Fourth-Generation SMR Developer was presented.348 The proposal was conditionally approved, with the rationale detailed and supplemented and the plan revised to proceed with the project after confirming economic viability through alternative methods of corporate valuation.349 However, as of the first half of 2025, the finalization of KHNP’s acquisition of a stake in TerraPower had not been confirmed. These developments have contributed to the high level of media interest in SMRs.

In fact, South Korea invested several hundred billion won (several hundred million U.S. dollars) over the past decades in the development of the 100-MW SMART (System-integrated Modular Advanced Reactor) design. A standard design certification for the original SMART reactor was completed in 2012, and a revised version, SMART100, received standard design approval from the Nuclear Safety and Security Commission in 2024.350 Despite these milestones and prolonged R&D efforts, to date, not a single SMART reactor has been sold and built domestically or abroad.

The nuclear power issue appeared in South Korea’s highly contentious political discourse following the declaration of emergency martial law by then-President Yoon on 3 December 2024. In a presidential address, delivered on 12 December after the martial law announcement, Yoon attempted to justify the unconstitutional measure by accusing the opposition of threatening democracy, national interests, and security through a wide range of traitorous, corrupt, and criminal actions, including that of undermining “national growth engines.” He stated:

The main opposition party is going so far as to shut off the Republic of Korea’s economic growth engines. This becomes evident when you look at the Democratic Party’s cutbacks in next year’s budget.

The financing for supporting the nuclear power ecosystem has been reduced. The funds allocated for the export of nuclear power plants to the Czech Republic have been slashed by 90 percent. The budget for developing next-generation nuclear plants has been virtually eliminated.351

However, these claims were demonstrably false. According to major media outlets’ analyses, the 2025 nuclear-related budget proposed by the MOTIE, consisting of 24 items totaling KRW488.9 billion (US$2024358.6 million), was passed without any reductions by the National Assembly on 10 December 2024. While there is no specific budget item titled “Czech nuclear export support,” the general nuclear export assistance program (Overseas Electricity Market Entry Support Project) received KRW18.3 billion (US$202413.4 million) and was approved without cuts.352

Yoon’s reference to “next-generation nuclear development” appears to relate to SMR-related funding. The KRW5.4 billion (~US$20244 million) allocated for the SMR Component Manufacturing Support Center and the KRW32.9 billion (US$202442million) for the i-SMR (Innovative SMR) R&D program were both passed in full. Separately, the MSIT also allocated KRW53 billion (US$202438.9 million) to its own innovative SMR development program, which likewise passed without amendment.353

The only nuclear-related item that saw a 90-percent reduction, as Yoon claimed, was not related to the APR-1000 model intended for export to the Czech Republic, rather, it was the Public-Private Partnership Program for Advanced Reactor Export Infrastructure, overseen by the MSIT, which focused on the basic design of an experimental next-generation Sodium-cooled Fast Reactor (SFR). This project’s budget was reduced from KRW7 billion (US$20245.1 million) to KRW0.7 billion (US$20240.5 million) after bipartisan agreement citing public controversy.354

Further, in June 2025, the Czech Minister of Industry and Trade stated:

The political situation in South Korea sometimes raised concerns in our country that it might jeopardize the nuclear tender. That is one of the reasons why, during our visit to South Korea in February [2025], we met not only with representatives of the government, but also with the then opposition Democratic Party. We were convinced that this was not the case. During meetings in both Korea and Prague, the then Korean opposition confirmed to me that it fully supports nuclear power and the project for new nuclear power plants in Dukovany.355

These developments illustrate that nuclear in general and SMRs in particular have become a highly visible and politically charged topic in South Korea, reaching into debates over constitutional legitimacy and the national budget.

The Czech Dukovany Nuclear Contract

WNISR2024 covered in some detail the KHNP/KEPCO–Westinghouse Part-810 export-control dispute—dismissed by a U.S. court on 18 September 2023, with an appeal filed on 16 October 2023—which formed a backdrop to the 2025 Dukovany contract.

On 4 June 2025, just one day after South Korea’s presidential election, the Czech government and KHNP signed the final contract for the construction of two nuclear reactors at the Dukovany site.356 The offer has been confirmed at around CZK200 billion (US$9.1 billion) per unit; though KHNP had earlier noted that “the final contract amount may vary depending on the outcome of the contract negotiations,” no updated amount was communicated.357 This marks South Korea’s second-ever nuclear reactor export—coming 16 years after the country’s first deal was signed with the UAE in 2009. Despite persistent efforts by the government and industry, no further exports had been secured until the Czech deal. Czech Prime Minister Petr Fiala personally announced the signing, calling it a “strategic investment” that would strengthen the country’s energy security. The timing of the contract—following the lifting of a court injunction and coinciding with Korea’s political transition—drew attention both domestically and internationally.358

The Czech nuclear project centers on building two APR-1000 reactors at the existing Dukovany nuclear power plant site and is aimed at replacing aging Soviet-era reactors and reducing the country’s dependence on coal and Russian energy technologies. In addition to KHNP, a consortium of Czech firms, including the national energy utility České Energetické Závody (ČEZ), will participate in project execution. The Czech government has emphasized the importance of partnering with a supplier from a democratic country that meets NATO and E.U. standards, a point that worked in favor of KHNP against competitors such as China’s CGN and Russia’s Rosatom, who were excluded early in the process.359

The path to the final contract was complex and extended over several years. In March 2022, the Czech government officially launched the tender for Dukovany Unit 5, with options to build Unit 6 as well. KHNP submitted its bid in November 2022, competing against France’s Électricité de France (EDF) and the Canadian-U.S. Westinghouse-Bechtel consortium. In July 2024, the Czech government confirmed KHNP as the preferred bidder, citing its competitive pricing, safety standards, and experience in exporting nuclear technology.360 This decision faced intense scrutiny, including legal and political objections from rival bidders (see past WNISR editions and section on the Czech Republic).

Legal hurdles escalated in May 2025, when a Czech court decided to temporarily forbid ČEZ to finalize the agreement with KHNP, pending its appraisal of the case brought by EDF against the Czech authority, which had twice dismissed the company’s claims of unlawful irregularities in the contract’s awarding process.361 A few days earlier, the European Commission had also pleaded to the Czech government for a halt to the signature of the contract, to allow for a review in response to a leaked document showing “significant indications” that KHNP may have received subsidies from the South Korean state “that are liable to distort the internal market”, in violation of E.U. regulations.362 However, on 4 June 2025, the Czech Supreme Administrative Court lifted the preliminary injunction imposed by the regional court, claiming notably that the Brno court had failed to take KHNP’s interest into account, thus, clearing the way for the contract to be signed.363 On the same day, the state-owned project company Elektrárna Dukovany II and KHNP formalized the agreement by signing an Engineering, Procurement and Construction (EPC) contract.

The Dukovany project is based on the APR-1000 design, a European-compliant variant of the APR-1400 reactor, which has already been built in South Korea and the UAE. Although KHNP has not yet constructed an APR-1000 domestically, the design has received certification from the European Utility Requirements (EUR) organization, strengthening its credibility for deployment in the European Union. The APR-1000 design includes additional safety features to the APR-1400 design such as double containment, a core catcher, and enhanced cooling systems to meet European safety standards.

Export Ambitions: The Risks of Low Bidding and the ‘On Time, On Budget’ Model

In 2009, KEPCO signed a US$20 billion contract as the lead of a consortium to build four APR-1400 nuclear reactors in the UAE.364 Under this contract, KHNP was responsible for reactor operations and technical provision, Doosan Enerbility for the main equipment, and Samsung C&T and Hyundai Engineering & Construction for construction. In March 2024, the fourth unit of the Barakah nuclear power plant began supplying electricity to the UAE grid, marking the full operation of the plant.

However, contrary to the often-used slogan ‘on time and on budget’ frequently cited by KEPCO and KHNP, the Barakah project experienced years of delays similar to many other new nuclear projects worldwide.365 For instance, Unit 1 was originally scheduled to begin commercial operation in 2017 according to a plan released in 2014,366 but it did not commence commercial operation until 2021367. Similarly, Unit 4 was delayed from its initial target of 2020 to 2024.

There has been no transparent, comprehensive report on the causes of these delays, but there are indications that the KEPCO-led consortium, including KHNP, may have been at least partly responsible. At the time of the UAE export deal, South Korea had not yet completed or operated a single APR-1400 reactor domestically. The Shin-Kori-3 (now Saeul-1), whose construction began in 2008, was the first APR-1400 reactor built in South Korea. In 2013, a major safety and corruption scandal erupted, and a government investigation reportedly found “277 faked certificates for parts used in 20 operating reactors, as well as 2,010 false documents at eight plants that were offline or under construction.”368 Furthermore, falsified test reports for control cables in several Korean reactors, including Shin-Kori-3 and -4, led to the replacement of over 900 kilometers of installed cable, which likely contributed to delays at both Shin-Kori and Barakah.369

Additional issues, such as coolant leakage from pressurizer safety valves in Shin-Kori-4 and voids found in the concrete containment buildings of Barakah Units 2 and 3, further complicated the timeline.370

In May 2025, it was reported that unresolved cost-assessment disputes between KEPCO and its own subsidiary KHNP over the Barakah project had escalated to international arbitration.371 On 7 May 2025, KHNP initiated arbitration proceedings at the London Court of International Arbitration over additional construction costs, claiming around US$1.1 billion according to KEPCO filings.372 Given that such arbitration processes typically take two to three years, both parties now face increased financial uncertainties.

Industry observers in South Korea point to the core cause of these disputes: the Korean export strategy based on fixed-price bidding, also known as the “On Time, On Budget” model.373 While attractive for winning contracts, this model is highly vulnerable to cost overruns and penalty risks when unforeseen construction delays occur.374 Large infrastructure projects are inherently subject to commodity price volatility and timeline extensions, increasing the likelihood of disputes. Because the nuclear contract terms were not publicly disclosed—KEPCO and the UAE signed a non-disclosure agreement (see WNISR2023)—and due to limited alternative sources of information on cost increases, it is difficult to assess how much construction costs have increased or how much actual profit has fallen. However, a delay of three to four years inevitably raises questions about who bears the added cost—KEPCO or KHNP, but in any case the Korean side. However, Hwang Joo-ho, president of KHNP, was quoted as saying that “when we built the Barakah NPP in the UAE, some of the increase in cost was due to the fact that we had to add new work (that was not in the contract) at the request of the client.”375

In fact, KEPCO reportedly has “consistently faced legal disputes with partners and subsidiaries” per Business Korea, including an arbitration case over additional costs that arose from project delays in the construction of Barakah filed by construction partners Hyundai E&C and Samsung C&T in July 2017,376 with a claim for an additional KRW600 billion (US$2017530 million). However, the arbitration resulted in a payout of only KRW20–30 billion (US$201917–26 million).377

EDF had also argued that the level of the fixed price for the Czech project—below €90/MWh (US$102.2/MWh)378—constitutes dumping and would only be possible because the Korean government subsidizes the cost gap. Korean nuclear industry officials rejected the assertion, attributing the cost advantage to Korea’s robust supply chain and project management capabilities.379

Since 2013, KEPCO has disclosed some cumulative revenue and profit data related to the UAE project in its annual business reports. Although these figures cover various domestic and international construction projects, they primarily reflect the Barakah project.380 KEPCO reported cumulative profit from the UAE project at KRW1.33 trillion (US$20191.1 billion) in 2019 (after the completion of Units 1 and 2), but this dropped to KRW72.1 billion (~US$202453 million) by the end of 2024. Given the reported total contract revenue of KR22.3 trillion (US$202416.35 billion), the implied cumulative profit margin stands at just 0.32 percent—a dramatic drop from the 10 percent margin initially anticipated in 2009.381

Moreover, the profit margin may decrease further or disappear altogether due to unresolved cost-sharing issues between KEPCO and KHNP. As of the end of 2024, KEPCO set aside KRW154.6 billion (US$2024113 million) as a “provision for liabilities” in its financial statement, indicating that cost disputes over the Barakah project are already reflected in its accounting. According to a media story citing KEPCO’s audit report, the company acknowledged that negotiations were ongoing with KHNP over extension-related costs and delay penalties and that a potential outflow of economic resources was expected. KEPCO noted that the amount recorded was based on auditor-approved figures and could change depending on future negotiations. If costs incurred during construction are not settled between KEPCO and KHNP—under the so-called “Team Korea”—KEPCO could reportedly face additional losses of up to KRW1.46 trillion (US$1 billion).382

South Korea’s low-price bidding strategy is often cited as a major factor behind diminishing returns. KEPCO’s US$20 billion bid for Barakah in 2009 was substantially lower than those from AREVA and GE-Hitachi, which played a decisive role in securing the contract. However, completing the project within such tight cost constraints—especially amid delays—makes significant profits unlikely if not leading to significant losses.

Adding to the challenge, Korean firms reportedly earned less because Westinghouse took over significant portions of the core equipment supply for the Barakah project. According to a report submitted to lawmaker Kim Sung-hwan’s office and seen by Korean-daily The Hankyoreh, Doosan supplied 51 percent of the primary system equipment (by value), while Westinghouse supplied 41 percent, and KEPCO E&C 7 percent.383 Toshiba, then-owner of Westinghouse, also supplied over half of the secondary systems (turbine generators). Westinghouse further received US$1.3 billion for licensing and technical support. Altogether, it is estimated that Westinghouse secured about 16 percent of the total project value, or around KR3.9 trillion (US$2.84 billion).384 An industry insider claimed that after KEPCO signed a cooperation agreement with Westinghouse in March 2010, Korean firms saw their share of the supply chain significantly reduced.385

Professor Park Jong-woon of Dongguk University’s Department of Energy and Electrical Engineering warned that the APR-1000 model planned for construction in the Czech Republic would require major design modifications compared to the APR-1400 used in Korea.

Prof. Park noted that “If South Korea, which has no experience in inland nuclear power plant construction,386 is forced to pursue ‘low-cost orders’ and ‘on-time delivery’ and fails to meet the completion period, it will suffer losses such as huge delay compensation.”387

KEPCO’s Continued Financial Crisis388

Korea Electric Power Corporation (KEPCO) was established on 1 January 1982 under the Korea Electric Power Corporation Act, with the purpose of developing electric power resources and conducting generation, transmission, transformation, distribution, and related businesses. On 1 April 2007, it was designated as a market-oriented public corporation under the Act on the Management of Public Institutions. KEPCO was listed on the Korea Exchange in August 1989 and on the New York Stock Exchange in October 1994.

As of the second quarter of 2025, KEPCO’s ownership structure remains majority-controlled by the South Korean government: the Korea Development Bank—wholly owned by the government—holds 32.9 percent, while the government directly holds 18.2 percent, together amounting to 51.1 percent.

WNISR covered KEPCO’s financial problems in its 2023 and 2024 editions and the crisis continues. KEPCO’s total debt at the end of 2024 stood at KRW120 trillion (US$202488 billion) on a standalone basis and at KRW205 trillion (US$2024150 billion) on a consolidated basis for KEPCO Group whose revenues stood at KRW93.4 trillion (US$202469 billion) and the profit level at KRW8.4 trillion (US$20246.2 billion).

In South Korea, which imports most of its fossil fuels, the share of coal, gas, and oil in power generation was 66 percent in 2009 and still accounted for close to 60 percent in 2024. This heavy reliance on imports means that when global fossil fuel prices rise and the government does not allow equivalent increases in electricity tariffs, KEPCO inevitably records operating losses. To cover these losses, the company issues bonds, leading to further growth in total debt. When international fossil fuel prices surged due to Russia’s invasion of Ukraine in 2022, KEPCO’s debt rose sharply. The Korean government ultimately raised electricity tariffs in 2023, and thereafter KEPCO’s debt level stabilized.389

According to KHNP’s filings, from 2009 to 2024 the company has never recorded an operating deficit. However, KHNP’s debt steadily increased from KRW13.5 trillion (US$200910.5 billion) in 2009 to KRW47 trillion (US$202434.5 billion) in 2024. With KEPCO Group’s extraordinary debt exceeding KRW200 trillion (US$145 billion) and KHNP’s debt exceeding KRW47 trillion (US$34.2 billion), the burden of financing the infrastructure investments required for the energy transition poses a significant challenge. Moreover, this massive debt load constrains KEPCO’s and KHNP’s ability to raise capital for any new investments or business ventures.

Taiwan Focus

About a decade ago, 2025 was designated as the target year for achieving Taiwan’s Nuclear-Free Homeland policy. As scheduled, Taiwan’s last operating reactor—Maanshan-2 at the Third Nuclear Power Plant in Hengchun, Pingtung—was closed on 17 May 2025. Beginning with that date, Taiwan officially entered a “nuclear-free” era.390

Compared to Germany, which completed its nuclear phaseout in 2023, the case of Taiwan attracted little attention from international media. However, like in Germany, challenges to the policy have emerged in Taiwan.

First, current domestic politics are unfavorable to the nuclear phaseout decision. Pro-nuclear opposition parties have secured a majority in the Legislative Yuan, the country’s parliament, and continue to advocate for restarting the operation of the Maanshan reactors. Some industrial leaders and economic advisors to President Lai have also publicly expressed support for nuclear energy.

Second, international and geopolitical factors have led to the perception of indirect costs of phasing out nuclear power. To reduce its trade deficit, the United States has pressured Taiwan to increase energy imports such as natural gas from the U.S., but it has also offered nuclear technology.

Third, public resistance to nuclear power has decreased, and confidence in nuclear energy has remained relatively high in recent years. At the same time, widespread biased information targeting renewable energy has significantly fueled public opposition to the construction of energy facilities such as offshore wind turbines or large solar plants.

Another critical challenge is the decommissioning of nuclear plants and construction of waste disposal facilities. The outdoor dry storage facility at the Second Nuclear Power Plant in Kuosheng, New Taipei City, was originally scheduled for completion in 2017, but construction was delayed because the Soil and Water Conservation Plan was not approved by the New Taipei City Government. In August 2024, the municipal government finally approved Taipower’s revised plan, and construction started on 31 December 2024. The removal of used fuel from the plant’s pool is expected to start as early as 2027 pending license approval.391 In addition, in May 2025, the dry storage facility was licensed for the Chinshan nuclear plant, also located in New Taipei City.392

In May 2024, the Ministry of Economic Affairs established the Office for Radioactive Waste Disposal.393 National Chengchi University Professor Tu Wen-ling was appointed as its director in April 2025; the office is tasked to release a legislative draft for a high-level radioactive waste disposal facility by the end of 2025.394

National Politics: Toward a Restart of Nuclear Plants?

After Democratic Progressive Party (DPP) candidate William Lai won the presidential election in January 2024, Chinese Nationalist Party (KMT) legislators renewed calls for scrapping the nuclear phaseout strategy.395 On assuming the presidency on 20 May 2024, Lai made a pledge to carry on the energy policy of President Tsai Ing-wen, his predecessor (see section on Energy Policy).

Soon after his appointment by Lai, Premier Cho Jung-tai reaffirmed that the new government “has no plans” to extend Maanshan’s operational lifetime.396 However, he has also indicated on several occasions the government’s openness to “new nuclear power” after 2030, if safety concerns associated with nuclear technology were resolved.397 A few weeks later, the National Climate Change Response Committee established under the direct authority of President Lai included both proponents and opponents of nuclear electricity.398

During the Committee’s first meeting, President Lai stated

Twenty-plus years ago, back when I was a member of the Legislative Yuan, working across party lines I co-sponsored a draft version of the Basic Environment Act. The Act has come to be known as Taiwan’s “environmental constitution,” and it is the first law in Taiwan to incorporate the concept of a “nuclear-free homeland.” Article 23 of the Act reads that the government shall formulate a plan to progressively achieve the goal of a nuclear-free homeland.

Although the DPP was the ruling party at that time within the executive branch, we did not have a legislative majority. The fact that we were able to get the Basic Environment Act passed by reaching a consensus between the ruling and opposition parties is proof that the concept of a nuclear-free homeland is not just an ideological stance of the DPP.399

Lai and the new cabinet continue however to face difficulties in stopping KMT legislators’ attempts to reverse the nuclear phaseout as DPP no longer holds a majority in the Legislative Yuan; the KMT is currently the biggest party by one seat over the DPP, and it has additional support from two independent lawmakers and the Taiwan People’s Party (TPP) with eight seats, making that a total of 62 seats for the opposition against 51 for the DPP.400

During Tsai Ing-wen’s eight-year administration, a fierce debate erupted between the DPP and KMT over several large-scale power outages, with the KMT blaming the government’s nuclear phaseout policy.401 Indicators like the System Average Interruption Duration Index (SAIDI) do not suggest any correlation with the closure of nuclear power plants. For example, between 2018 and 2020, the SAIDI was higher than in 2023 and 2024.402 Nevertheless, in July 2024 (still under the Tsai administration), the KMT proposed an amendment to the Nuclear Reactor Facilities Regulation Act to extend the operating lifetime of nuclear reactors by 20 years to 60 years, but the amendment did not garner sufficient support during the standing committee’s review.403

The KMT has found a new window of opportunity with its current majority in the Legislative Yuan. In March 2025, KMT and TPP lawmakers jointly proposed draft amendments to relax the deadline for license renewal applications and to extend the reactor operational limit to 60 years, theoretically paving the way for restarting the Maanshan plant.404 In response to a KMT legislator’s question, Minister of Economic Affairs Kuo Jyh-huei stated that Taiwan would need to acquire new fuel elements to continue operations, a process that would take at least 16 to 18 months.405 Although DPP legislators repeatedly raised concerns over unresolved nuclear waste and safety issues, they were ultimately outnumbered by the opposition,406 and the amendment passed on 13 May 2025—just four days before the scheduled closure of Maanshan Unit 2.407 The actual implementation of this amendment still requires additional procedures and safety checks proposed by Taipower, and no final decision has been made regarding the need for further Environmental Impact Assessments (EIAs).

The proposed restart also raised another issue rooted in the legacy of nuclear development in Taiwan. As the retiring reactors were constructed under the KMT’s authoritarian regime prior to the 1980s, they were exempt from EIAs. As a result, there is no legal precedent for how to conduct an EIA for the reactivation of closed reactors. The day after the amendment passed, Minister of Environment Peng Chi-ming stated before the Legislative Yuan’s Committee on Social Welfare and Environmental Hygiene that he found no existing legal framework to guide such reactivation. DPP legislator Lin Shu-fen supplemented that reactivation should follow the existing regulations for infrastructure development projects; consequently, not only is it required to conduct an evaluation of the cost, but also an environmental impact assessment.408 Since the plant’s license has expired and there is no direct capacity for operation, if a restart is requested, inspections and evaluations will be carried out but the Nuclear Safety Commission could reject the restart application for safety concerns.409

Pro-nuclear lobbying had experienced a major setback in December 2021, when a proposal to resume the construction of two reactors at the Fourth Nuclear Power Plant in Lungmen was rejected in a referendum, indicating popular support for a nuclear-free policy at the time.410 Regardless of the vote’s outcome though, the plant was unlikely to be completed or become operational due to the dire state of the project (see past WNISR editions).

More recently, some business leaders and KMT politicians have promoted the idea of deploying Small Modular Reactors (SMRs) “in every administrative region of Taiwan.”411 Environmental organizations such as Green Citizens’ Action Alliance (GCAA) and Citizen of the Earth, Taiwan, have argued that in addition to further worsening the problem of nuclear waste,412 SMRs would be too slow and too expensive to efficiently address the climate crisis.413

The latest initiative from the KMT–TPP coalition to reverse the nuclear phaseout is a referendum scheduled for 23 August 2025 on the following question: “Do you agree that the Third Nuclear Power Plant [Maanshan] should continue operating, provided that the competent authority confirms there are no safety concerns?”414 The question is confusing: since the plant has already been closed, it cannot “continue operating”; it can only be restarted. On the same day another vote is to be held to decide on the recall of seven KMT lawmakers, following a first vote on 26 July 2025 for 24 KMT lawmakers. This recall movement results from the national mobilization of civil society organizations, started in May 2024, to reverse the balance of power at the Legislative Yuan on the claim that the KMT-TPP coalition is threatening the island’s sovereignty by making excessive concessions to Beijing.415

International Political Pressure

Another major factor influencing Taiwan’s nuclear phaseout policy is its diplomatic environment and geopolitical concerns, particularly the relationship with its most important ally, the United States. Following his inauguration, U.S. President Donald Trump swiftly initiated a trade war by imposing tariffs on many countries. In response, the U.S.’ East Asian allies—Japan, South Korea, and Taiwan—swiftly signed agreements to access liquefied natural gas (LNG) from a future project in Alaska.416 On 20 March 2025, Raymond F. Greene, the director of the American Institute in Taiwan (AIT, the de facto embassy), mentioned that the U.S. could provide Taiwan with “the full range of energy solutions from geothermal to nuclear to advanced grid technologies” and declared that the Trump administration was “especially eager to share […] LNG resources with close partners.”417

Tung Tzu-hsien, a pro-nuclear businessman and member of President Lai’s National Climate Change Committee, quickly reacted to Greene’s remarks, arguing that Taiwan should revise its nuclear regulatory laws to accommodate the expansion of its energy portfolio.418

Regarding potential products, AIT officials provided a comprehensive list without limiting it to specific items. In an interview, Greene added SMRs and services related to nuclear waste management to his proposed energy package.419 Although the policy remains ambiguous, Lai’s government’s reluctance is palpable. On 17 April 2025, KMT legislator Chen Ching-hui questioned Environment Minister Peng Chi-ming about the possibility of Taiwan purchasing SMRs from the U.S. as a “bargaining chip” in tariff negotiations. Peng responded that he had not been involved in such discussions but emphasized the high cost, potential supply chain issues, and technological challenges of SMRs, all of which require public dialogue. He also cautioned that Taiwan’s aging nuclear plants lack environmental impact assessments, raising safety concerns. Peng reiterated the government’s openness to discussion but stressed that Taiwan lacks the regulatory mechanisms, technical capacity, and social consensus for restarting nuclear power, calling it a complex and lengthy process.420

Energy Policy

Public opposition to nuclear power in Taiwan reached a new high immediately after the Fukushima disaster was triggered in March 2011.421 Sending shock waves through Taiwan, the event increased public acceptance for a nuclear phaseout and an energy transition. Having returned to power in 2016, the DPP announced the New Energy Policy aimed at establishing “a low carbon, sustainable, stable, high-quality and economically efficient energy system” through an energy transition and energy industry reform.422 On 12 January 2017, the Electricity Act Amendment completed its third reading in the legislature, ushering Taiwan’s energy transition, including the nuclear phaseout.423

During the second term of Tsai’s presidency (2020–24), Taiwan aspired to become a leading country of renewable energy in the Asia-Pacific region.424 The island’s renewable energy potential is significant, and in 2021, the Global Wind Energy Council estimated Taiwan’s offshore wind technical potential to be as high as 494 GW.425 In 2019, the government set the goal to install 5.7 GW of offshore wind capacity by the end of 2025.426 In May 2021, the target was increased to 15 GW with annual deployment of 1.5 GW over the decade. The target has been confirmed in 2023.427 Nevertheless, targets for end-2025 placed solar capacity at 20 GW and combined renewable energy generation at 20 percent of the power mix.428

These goals remained ambitious, especially since until a significant boost in 2022, the development of renewable energy had been slow. In 2022, the installed renewable energy capacity reached 14.1 GW, representing 23 percent of installed capacity, and renewable power generation increased by 36 percent to reach 23.9 TWh, contributing 8.3 percent of the total electricity production. The year 2022 marked the first time that the share of renewable energy—including geothermal, biomass, waste, and hydro—in total electricity production slightly surpassed nuclear power (8.3 percent versus 8.2 percent), with a 163-percent increase in offshore wind power generation compared to the previous year and a 57-percent increase in total wind production.429

The trend continued in 2024 with the share of renewable energy in total electricity production reaching 11.7 percent—compared to nuclear power’s 4.2-percent contribution—mainly generated by solar at 5 percent and wind at 4 percent (onshore and offshore combined). Solar production was up 18.5 percent while wind power generation jumped 68.5 percent compared to the previous year. The installed capacity of renewable energies reached 21 GW up from 18 GW in 2023.430 The deployment acceleration has also been noted by investors. Taiwan again moved up two places in the Ernst & Young’s “Renewable Energy Country Attractiveness Index 2023” to rank 24th in 2023,431 and it maintained that position in 2024.432

However, as of mid-2025, Taiwan remains a fair step away from its goals—solar PV capacity is at 14.8 GW, wind at 4 GW, and the contribution of renewables to total power generation stands at 13 percent,433 compared to the respective targets of 20 GW, 4.7 GW and 20 percent. The country remained heavily dependent on imported fossil fuels in 2024 with an overall primary energy dependence rate of 95.2 percent. In electricity generation, LNG and coal combined made up over 80 percent in 2024. LNG-generated electricity shot up from less than 40 TWh in 2004 (and less than 20 TWh in 2000) to 122.5 TWh in 2024, contributing the largest portion of total electricity generation at 42.4 percent, now above coal’s 39.2 percent (see Figure 41).434

1. Electricity Production in Taiwan, 2000–2024

Source: MOEA, Energy Handbook, Various Years

Despite not being able to participate in the Paris Agreement and COP negotiations, in April 2021 the Taiwanese government unilaterally pledged to achieve Net-Zero by 2050, announcing that it would draft regulations to that end and accelerate the implementation of existing targets.435 In 2023, Taiwan promulgated the Climate Change Response Act, which replaced the former Greenhouse Gas Reduction and Management Act of 2015, making legally binding the goal of Net-Zero by 2050.436 On 29 August 2024, Taiwan’s Ministry of Environment finalized three regulations for the carbon tax system, officially launching the carbon pricing era. Carbon fees will begin in 2025, based on that year’s emissions, with payments collected from 2026 onward. The system allows emitters time to plan and meet 2030 reduction targets through voluntary reduction plans.437

Per capita energy consumption has hardly gone down over the past decade, while per capita electricity consumption increased by 14 percent over the decade 2013–2022, decreased slightly by 1.2 percent in 2023, and went back up in 2024 by 2.6 percent year-on-year.438 Peak load experienced the strongest growth rate at 19.5 percent over that decade to exceed 40 GW for the first time in 2022.439 Hitting a new record in July 2024 at 41.4 GW, it has since remained just below that, reaching 40.9 GW on 2 July 2025.440

The previous government’s strategy aimed to “promote green energy, increase natural gas, reduce coal-fired power, and achieve nuclear-free.”441 This implies that Taiwan would have continued to see a substantial increase in natural gas consumption aiming to provide 50 percent of gross electricity production in 2025, compared to 42.4 percent achieved in 2024.442

In March 2022, Taiwan’s Pathway to Net-Zero Emissions in 2050 was unveiled.443 The strategy was based on a NT$900 billion (US$202232.4 billion) budget to 2030, of which NT$210.7 billion (~US$20227.6 billion) were allocated to “renewables and hydrogen”, and another NT$207.8 billion (~US$20227.5 billion) were to be invested in “grid and energy storage”. The associated goals included a total of 40 GW of combined wind and solar capacity by 2030, and by 2050, an installed capacity of 40–80 GW in solar and 40–55 GW of offshore wind, for a combined share of more than 60 percent.

Under the presidency of William Lai, the new DPP government is expected to follow the preceding administration’s footsteps for the energy transition, but tensions in national and international politics might render implementation more challenging. Another problem is the gap between the reality of nuclear energy and its public representations as shown in a series of surveys such as Academia Sinica’s Taiwan Social Change Survey Programme.444

Public Attitudes and Misinformation

The Fukushima disaster of 2011 had a strong impact on public opinion in Taiwan.445 However, according to a 2024 analysis, in recent years, the percentage of respondents who agreed or strongly agreed that nuclear power plants posed a high risk which was nearing 80 percent in 2017 dropped to 64.3 percent in 2019, 66.5 percent in 2021, and 62.6 percent in 2022.446

Meanwhile, although in 2017, the share of nuclear power had already fallen to less than 10 percent—the first time in over four decades—a large proportion of citizens believed that nuclear energy was Taiwan’s primary source of electricity. In a 2018 survey conducted by the Risk Society and Policy Research Center, National Taiwan University, although 82 percent of respondents expressed concern about the development of Taiwan’s energy policy, only 32 percent correctly identified coal-fired power as Taiwan’s primary source of electricity and 44 percent believed that nuclear power contributed the largest share.447 This view persists: in a poll conducted by Pearson on behalf of media portal Tai Sounds, in April 2024 more than half of respondents (53.4 percent) still considered nuclear power one of Taiwan’s top three sources of electricity.448 Another survey conducted at the end of 2024, highlighted a strong risk perception of global warming, yet fewer than 20 percent of respondents correctly identified nuclear power’s low share (below 10 percent), while nearly 25 percent of college-educated respondents believed it accounted for over 51 percent of Taiwan’s electricity mix.449

The general misconception regarding nuclear, partly rooted in education during the peak of nuclear energy, has led to an overestimation of nuclear power’s significance in the supply mix. Misinformation from pro-nuclear advocates and the media has further undermined public confidence in Taiwan’s nuclear-free policy,450 leaving public opinion evenly split (~50 percent support/opposition). According to research projects supported by Academia Sinica from 2017 to 2024, notably, women, less-educated citizens, and the DPP supporters were more likely to favor a nuclear-free homeland, while men, higher-educated individuals, and the KMT supporters prioritized decarbonization and energy security over nuclear risks—resulting in a prevailing sense of “nuclear-free anxiety.”451

Misperceptions about nuclear waste management are also widespread. According to the 2019 Taiwan Social Change Survey (Round 7) led by researchers at Academia Sinica, on behalf of the National Science Council, which focused on technology and risk, only less than 12 percent of respondents were aware that Taiwan’s high-level radioactive nuclear waste (spent fuel) was currently stored at nuclear power plants.452 Interestingly, the survey also found that supporters of the statement “we shall keep using nuclear energy,”—53 percent in this survey—did not demonstrate better knowledge of nuclear waste management than those in favor of a nuclear phaseout; in both categories, only 14 percent had the right answer.453

Taiwan has recently been targeted by foreign disinformation campaigns aimed at undermining its energy policy. As a recent study on climate and energy misinformation in Taiwan highlights, disinformation about nuclear power has become a significant force shaping public discourse, particularly amid Taiwan’s efforts to phase out nuclear energy under the 2025 Nuclear-Free Homeland Plan.454

Reactor Closures and Spent-Fuel Management

As reported in previous WNISR editions, Taipower announced the closure of Chinshan-1 on 5 December 2018. The reactor had not generated power since the end of 2014. Chinshan-2 had been off-grid since June 2017, but was not officially closed until 15 July 2019, when its 40-year operating license expired.

On 1 July 2021, Taipower stated that due to lack of spent-fuel storage-capacity, Kuosheng-1 was closed six months ahead of schedule.455 The closure of the reactor, located on the northern coast, only 22 km away from Taipei, was originally slated for 27 December 2021, the day its operating license expired.456 Kuosheng-2 ceased operating on 14 March 2023.

Finally, the remaining two PWRs at Maanshan closed on 26 July 2024 and 17 May 2025, respectively. In line with the nuclear phaseout and current regulation, the application to disconnect the units from the grid had been submitted in July 2021.457 (See Figure 42.)

1. Histogram of Taiwan Nuclear Fleet

Sources: WNISR with IAEA-PRIS, 2025

Despite the fact that the spent-fuel pools of the Chinshan and Kuosheng power plants have been filling up, the government made little effort to arrange dry cask storage for the high-level radioactive waste and little more than 1 percent of the fuel has been taken out of the pools (see Table 10). Successive governments paid little attention to intermediate storage and final disposal of spent fuel.458 In 2023, GCAA and other environmental groups lobbied in vain to legislate on the issue.459

1. Spent Fuel at Taiwan’s Nuclear Power Plant Sites

Site / Unit	Grid Connection/Closure	Spent Fuel at Taiwan’s Nuclear Sites (in Numbers of Fuel Assembly)
		Reactor Core	Spent-Fuel Pool	Dry Storage Facility	Total
Chinshan-1	1977/2014	54	3,056	280	6,874
Chinshan-2	1978/2017	408	3,076
Kuosheng-1	1981/2021	624	4,808	NA	10,868
Kuosheng-2	1982/2023	624	4,812
Maanshan-1	1984/2024	0	1,879	NA	3,785
Maanshan-2	1985/2025	0	1,906

Source: IAEA-PRIS (grid connection years) and Nuclear Safety Commission, 2025460

Reactor closures, in general, have not ushered in the actual technical decommissioning phase, the bulk of which would come only after the spent-fuel assemblies have been unloaded from the cores and pools and placed in dry cask facilities.

The operating license for Unit 1 of the First Nuclear Power Plant (Chinshan) expired in December 2018, leading to its official closure. To prepare for decommissioning of Chinshan-1, Taipower proposed a plan to establish a dry cask storage facility, adopting a system certified by the U.S. Nuclear Regulatory Commission (NRC). The associated Soil and Water Conservation Plan was approved by the New Taipei City Government in 2010 and completed in 2013. However, following the completion, Taipower submitted 13 applications for subsequent detailed design modifications for review, all of which were rejected by the government of New Taipei City.461 In April 2014, the city’s government formally rejected the entire project and refused to grant approval. During this period, New Taipei City was governed by Mayor Eric Chu, KMT, who is current chairman of the party.462

Following a decade long legal battle, on 16 March 2023, the Taipei High Administrative Court ruled in favor of Taipower, ordering the New Taipei City Government to approve Taipower’s dry storage plan.463 The municipal government ultimately approved the dry cask storage facility at Chinshan in May 2024,464 followed by the approval of the dry cask storage facility at Kuosheng in August 2024.465 (See Taiwan in Decommissioning Status Report).

In 2023, a new independent body, the Nuclear Safety Commission (NSC), replaced the former nuclear regulator, the Atomic Energy Commission (AEC).466 While scaled down from second-level to a third-level agency, the NSC shifted focus from nuclear development to monitoring and regulating nuclear decommissioning, as stipulated by the Nuclear Safety Commission Organization Act, the legal framework for NSC’s establishment.467 According to the Act, the new commission shall oversee and implement waste management, which will be a major challenge in the coming decades.

Ukraine Focus

For the first time, a country with a developed nuclear power sector has been subjected to a full-scale military attack and has remained at war for more than three years. Russia’s full-scale invasion of Ukraine in February 2022—the largest military conflict in Europe since World War II—has created fundamentally new challenges for the international nuclear safety and security system. Previously unthinkable scenarios have become routine: shelling of operating nuclear power plants and nuclear facilities, disruptions in power and water supply to nuclear reactors, and even the seizure and occupation of nuclear sites. Throughout the war, however, nuclear power plants have remained the backbone of Ukraine’s electricity system.

Over the past 10 years—both before and after the 2022-invasion—nuclear power has consistently accounted for around 50 percent of Ukraine’s total electricity generation. However, due to partial destruction of the energy infrastructure and a decline in consumption, the country’s overall electricity output dropped by around 35 percent between 2021 and 2024. In 2024, Ukrainian reactors produced only 53 TWh (gross)—about the same as in 2023 but almost 40 percent lower than the pre-war level of 86.2 TWh468 in 2021—primarily due to the seizure and shutdown of the six reactors at the Zaporizhzhia nuclear plant.

Ukraine has 15 operable reactors (two VVER-440s and 13 VVER-1000s) with a total installed capacity of 13.1 GW. These include six units at Zaporizhzhia that, since March 2022, have been occupied by the Russian army assisted by technical experts from Rosatom. The six reactors have been shut down since September 2022 and are considered in LTO. The country also has four closed RBMK reactors at the Chornobyl nuclear power plant, including Unit 4, which underwent a disastrous accident in 1986.

As of 1 July 2025, the mean age of Ukraine’s reactor fleet, including reactors in LTO, is 36.4 years. Only three reactors are younger than 30 years. All others have been granted lifetime extensions for 10 or 20 years, as described in WNISR2024.

Newbuild Projects

In April 2023, the Cabinet of Ministers approved a new Energy Strategy for Ukraine until 2050.469 In December of the same year, at the COP28 conference, Ukraine presented the report Clean Energy Roadmap: From Reconstruction to Decarbonization in Ukraine.470 Linked to the goals of Ukraine’s energy strategy and developed with the support of the United States, the report presented scenarios for the development of a low-carbon energy system in Ukraine. According to the proposed scenarios, the country’s nuclear energy production should increase by at least 2.5 times by 2050 compared to the current level, accounting for between one-third and 60 percent of electricity generation, depending on the scenario, and up to half of total primary energy production. The remainder should come almost entirely from renewable energy sources.

In January 2024, Ukraine’s then-Minister of Energy German Galushchenko471 announced that the country would begin construction during the year of four nuclear units at Khmelnytskyi (also spelled Khmelnitsky, including by the IAEA), where two reactors are already operating.472 As described in past WNISR editions, Ukraine has a long history of attempting to complete two abandoned VVER-1000 reactors—Units 3 and 4 of the Khmelnytskyi nuclear power plant—whose construction originally started nearly 40 years ago. These efforts were revived in response to the loss and destruction of a significant share of Ukraine’s generating capacity following Russia’s full-scale invasion. According to Galushchenko, “Completing the construction of the Khmelnytskyi NPP [Nuclear Power Plant] is a strategic task for Ukraine.”473

However, the feasibility of completing these units largely depends on the potential purchase of key equipment—primarily reactor vessels and steam generators, originally intended for Bulgaria’s unfinished Russian-designed Belene nuclear power plant. In July 2023, the National Assembly of Bulgaria mandated the Ministry of Energy to enter into negotiations with Ukraine but requested not to sell the equipment below the initial purchase price of ~€600 million (US$2023649 million).474 Discussions dragged on, and on 11 September 2024, the Bulgarian parliament granted a 180-day extension of the negotiation deadline.475

On 11 February 2025, the Verkhovna Rada of Ukraine approved a draft law allowing the Ukrainian nuclear operator Energoatom to purchase equipment from Bulgaria for the completion of the two units at Khmelnytskyi.476 The law was signed by President Volodymyr Zelenskyy on 13 March 2025. According to proponents of the project, this purchase represents the fastest way to acquire two large nuclear units, which would significantly reduce power shortages in the country. They claim the reactors could be completed within three to four years.477 The Ministry of Energy had indicated in February 2025, that construction work was 80 percent complete at Unit 3 and 25 percent at Unit 4.478

However, critics of the project, as cited in WNISR2024, emphasized concerns that remain relevant a year later. They noted that the equipment is not complete and that, so far, no agreements to manufacture the missing equipment exist. They also pointed out that there was no law authorizing the construction of Units 3 and 4, no design, no safety report, and no construction license. This remains valid as of mid-2025.

In February 2025, Energoatom commissioned an update of the project’s feasibility study.479 The work was supposed to be completed within four months (by early June), and the Ministry of Energy called on Energoatom to accelerate this work in accordance with the decision of the Verkhovna Rada,480 however no public information on the results had been released as of 1 July 2025.

In April 2025, senior officials in the Bulgarian government announced that the country would not proceed with the sale of Belene reactor equipment to Ukraine.481 The decision was reportedly driven by a renewed interest in keeping the equipment for potential use at Bulgaria’s own nuclear facilities, such as Kozloduy. Given Bulgaria’s political instability and history of inconsistent nuclear policy, the situation remains fluid; the decision may still be revised, though there is also a possibility that the deal will not move forward.

According to the IAEA, Khmelnytskyi-3 and -4 are officially under construction, but WNISR removed them from the construction list as no active work has been reported in some four decades despite all attempts to revive the project. The last time the government attempted to restart construction of the plant, the total cost of completing Khmelnytskyi Units 3 and 4 was estimated at UAH201772.3 billion (~US$20172.7 billion).482 But there has been no independent evaluation of the cost estimate.

In June 2022 Energoatom and Westinghouse signed an agreement to construct up to nine AP-1000 units in Ukraine.483 In January 2023, the Cabinet of Ministers approved the launch of a feasibility study for constructing two such reactors—Khmelnytskyi-5 and -6—by 2032, with an expectation that they would cost around US$5 billion each.484 The proposed timeline and cost seem excessively optimistic, especially considering that the forerunner Vogtle plant in the U.S. took over a decade to complete at more than triple the cost projected for Ukraine.

In April 2024, it was announced that a “symbolic cubic meter” of concrete for Unit 5 had been poured at Khmelnytskyi in the presence of U.S. and Ukrainian officials as well as the CEOs of Westinghouse and Energoatom.485 Bridget Brink, then-U.S. Ambassador to Ukraine,486 reiterated at the ceremony that these units would be the first of nine AP-1000 reactors to be built in Ukraine in cooperation with Westinghouse.

But similar to Units 3 and 4, work on Khmelnytskyi Units 5 and 6 remains effectively inactive with no formal feasibility study, safety report, or construction license. The symbolic ceremony held in April 2024 cannot be considered the actual start of construction, as it involved concreting of a drainage channel rather than the nuclear island itself.487 Thus, by mid-2025, despite the stakeholders’ intentions and high-profile announcements, not a single new unit can be considered in active construction.

Further, Energoatom has shown interest in deploying Small Modular Reactors (SMRs) and entered cooperation agreements with various potential technology providers and partners, such as NuScale, Holtec, Rolls-Royce, and Hyundai E&C, since 2021.488 In September 2023, Energoatom and Westinghouse signed a Memorandum of Understanding (MoU) on the development and deployment of an AP300, a yet-uncertified SMR design. The MoU established a joint working group to develop licensing, contracting, and local supply chains. Westinghouse is hoping that design certification could take place by 2027 and construction could start in 2030.489 As Westinghouse has not even submitted an official application for design certification to the U.S. regulator, the timeline appears optimistic, even under new government pressure on the regulator to accelerate procedures.490

Reducing Russia Dependencies

Ukraine is the VVER-reactor country that has advanced the furthest in moving away from dependence on Russian nuclear technologies and supplies (see Annex 3: Fuel Supply for Soviet-designed Reactors in the E.U. and Ukraine). In 2001, Westinghouse began the development of VVER-1000 fuel manufacturing following a government agreement between Ukraine and the U.S. on energy cooperation and safety. In 2008, Energoatom and Westinghouse signed the first contract for fuel delivery; in 2018, the contract was extended to 2025.491

By the time Russia invaded it in 2022, Ukraine had already transitioned six of its VVER reactors partially or fully to Westinghouse fuel. Four of those reactors were at the Zaporizhzhia site. In June 2022, the contract with Westinghouse was expanded to cover all operational Ukrainian units, stipulating a full transition to Westinghouse fuel.492

In September 2023, the first VVER-440 fuel assemblies delivered by Westinghouse were loaded at the Rivne plant,493 and in March 2024, VVER-1000 fuel assemblies were delivered to Khmelnytskyi494.

After terminating its contracts with Russia’s Rosatom, Ukraine’s Energoatom signed agreements with Western suppliers covering the entire nuclear fuel supply chain. In 2023, Canada’s Cameco was contracted to supply natural uranium and conversion services for Ukraine’s entire nuclear fleet through 2035.495 In autumn 2023, a contract was signed with the British-Dutch-German company Urenco for uranium enrichment services until 2035, with an extension option through 2043.496 In March 2025, a similar agreement was signed with France’s Orano, covering services through 2040.497

Ukraine has demonstrated that even under wartime conditions, it is possible to transition from Russian supplies to Western sources while keeping 60 percent of its nuclear fleet operational. This experience is highly relevant for all countries relying on Russian nuclear supplies but are aiming to transition to other sources (see Russia Nuclear Interdependencies).

Power Sector Under War Conditions

Since the beginning of the full-scale Russian invasion, Ukraine’s electricity sector has faced major disruptions. According to a United Nations Development Programme (UNDP) report,498 in the early months of the invasion significant degradation was caused by the temporary occupation of parts of the country, the reduction in industrial activity, and the relocation of millions of people to safer regions or abroad. The situation worsened in October 2022, according to the report, when Russia began systematic missile and drone attacks on Ukraine’s energy infrastructure. These attacks left around 12 million people without electricity, disrupted water and heating supplies, and caused widespread hardship during sub-zero winter conditions.

According to the UNDP assessment, in the first 14 months since the start of the invasion, Ukraine’s total operational generation capacity decreased by more than half—from 37.6 GW at the beginning of 2022 to 18.3 GW as of 30 April 2023. The analysis found that

At the end of April 2023, there was not a single TPP [Thermal Power Plant] or HPP [Hydroelectric Power Plant] in Ukraine that was not damaged to some extent due to military activities and missile attacks on energy infrastructure facilities. (…) In the transmission network, 42 out of 94 crucial high-voltage transformers have been damaged or destroyed.499

Due to the complete shutdown of Zaporizhzhia, operable nuclear capacity declined by 44 percent from 13.8 GW to 7.7 GW. Available hydro capacity dropped by 29 percent from 6.6 GW to 4.7 GW and renewable capacity decreased by 24 percent from 8.1 GW to 6.2 GW, “as a large number of wind and solar power plants located in southern Ukraine are currently in areas under the temporary military control of the Russian Federation, are damaged or are in the combat zone,” the UNDP assessment found. For the mid-term, the UNDP analysts see a chance to accelerate the energy transition as “the loss of obsolete coal-fired generation opens an opportunity for their replacement using greener alternatives and decentralizing generation capacities.”500

According to a joint report by the World Bank, the Ukrainian government, the European Commission, and the U.N., in the course of 2024, Ukraine’s energy infrastructure sustained additional damage due to 13 massive, targeted attacks. As of 31 December 2024, despite improved preparedness for attacks compared to the previous year and the restoration of several gigawatts of capacity, the overall installed generation capacity available for use dropped to 15 GW, which is significantly lower than the 18–19 GW needed for peak demand, especially in winter.501

Total electricity generation in 2024 dropped to 101 TWh, which is about 3 percent less than the previous year and less than two thirds of the pre-war year 2021 (156.5 TWh).502

Under such conditions, electricity imports from E.U. countries have become extremely important, especially for balancing the power system. In March 2022, Ukraine disconnected its power grid from Russia and subsequently synchronized with the E.U. grid.503 Starting in 2023, Ukraine shifted from being a net exporter to a net importer of electricity, importing a record 4.4 TWh in 2024—mainly from Hungary and Slovakia.504 The growth in imports was also made possible by infrastructure repairs, which increased import capacity from 1.7 GW to 2.1 GW by December 2024,505 though this was still insufficient to make up for the Ukrainian power deficit of almost 2.3 GW in summer 2024506 and an estimated 3–4 GW during peak winter demand.

Up to November 2024, the G7+ Ukraine Energy Coordination Group says it “mobilized more than US$ 5 billion in support of Ukraine’s energy sector.”507 International support for Ukraine’s energy sector focuses on restoring infrastructure, expanding decentralized renewables, and strengthening cross-border links—aligned with the EU Clean Energy Package and climate goals. It also emphasizes better protection of facilities to reduce humanitarian and environmental risks.

At the same time, according to estimates from a joint assessment by the World Bank, the Ukrainian government, the European Commission, and the United Nations, as of the end of 2024, the amount needed to restore Ukraine’s energy sector is about US$68 billion.508

Russian Attacks on Nuclear Facilities

On the first day of Russia’s invasion of Ukraine on 24 February 2022, Russian troops entered the Chornobyl Exclusion Zone from Belarus, advancing along the shortest route toward Kyiv, and seized the Chornobyl nuclear power plant. By 31 March 2022, Russian forces had withdrawn from the area509 as part of a large-scale retreat following the failed attempt to capture the Ukrainian capital.

A week after the invasion began, Russian forces from the southern group, advancing from Crimea (occupied since 2014), approached Europe’s largest nuclear power plant, Zaporizhzhia, and on 4 March 2022, they seized the facility.510 According to the IAEA, representatives of State Atomic Energy Corporation Rosatom arrived at the Zaporizhzhia site “a few days after the Russian military took control.”511 The IAEA also noted that within a few months an “increasing number of Russian technical staff” from Rosatom and its subsidiary Rosenergoatom “has been observed at the site.”

In August 2022, the IAEA reported the first serious shelling of the plant;512 since then, such attacks have continued intermittently. For more than three years now, the plant has been located directly on the frontlines, which run along the banks of the Kakhovka Reservoir (which was destroyed in the summer of 2023).

On 1 September 2022, IAEA Director General Rafael Grossi visited the station for the first time during the war and established the presence of the IAEA Support and Assistance Mission to Zaporizhzhya (ISAMZ) at the facility.513 Since then, representatives of the agency have maintained a continuous presence at the site. However, the IAEA complained that the Mission “continued to face some restrictions in obtaining timely and appropriate access to all areas of relevance to nuclear safety and security and in having open discussions with all relevant staff” at the site.514

In September 2022, the remaining operating reactor units were put in cold shutdown mode.

The same month, President Putin formally declared that the regions of Luhansk, Donetsk, Kherson, and Zaporizhzhia were part of Russia.515 In October 2022, Putin signed a decree that transferred the power plant to Russian ownership with station management being entrusted to Rosenergoatom. Rosenergoatom established a Joint Stock Company “Operating Organization of Zaporizhzhia NPP”, to operate the plant.516 The IAEA has acknowledged that it has no authority to enforce any of the resolutions that have been passed by the United Nations calling on Russia to withdraw from the power plant.517

The most significant threats to radiation and nuclear safety during the occupation of the Zaporizhzhia plant have been linked to the following safety and security risks (for details and further references, see Nuclear Power and War in WNISR2022):

• The loss of power necessary to ensure the operation of active safety systems—primarily for cooling nuclear fuel in the reactor cores and storage facilities—to prevent a meltdown accident. Since the end of 2022, the situation with Zaporizhzhia’s grid connection has remained fragile—qualified as “extremely vulnerable” by IAEA’s Rafael Grossi in mid-June 2025518—and virtually unchanged: with periodic outages, it has been connected to the Ukrainian power system by only one or two lines compared to ten before the occupation. On 4 July 2025, the station was once again (for the ninth time since the occupation began) completely disconnected from the power grid, relying solely on backup diesel generators for over three and a half hours.519 In the case of an operating reactor, a station blackout followed by the failure of emergency power generators to start up almost certainly leads to a meltdown accident within about one hour.
• The loss of water supply needed to cool the nuclear fuel inside the reactor cores and in the spent fuel pools. At Zaporizhzhia, this risk exacerbated after the destruction of the Russian-controlled Kakhovka Hydroelectric Power Plant dam and the draining of the Kakhovka Reservoir.520 For Zaporizhzhia, this meant the loss of the ultimate heat sink. However, given that all reactors at the plant are shut down, the demand for water is significantly lower than during normal operation, and the loss of the reservoir does not pose an immediate threat to the plant.521 Later, the IAEA communicated that 11 dedicated wells were drilled on the plant’s territory to improve water availability for cooling purposes.522 Nevertheless, the water access situation remains volatile.
• Deliberate attacks or accidental shelling of Zaporizhzhia as well as the other nuclear power plant sites. Russia has been actively fortifying the Zaporizhzhia site, mining its perimeter, digging trenches, and stationing troops both in and around the facility—effectively turning it into a fortified military position.523 The facility is also being used to shield Russian firing positions and to launch attacks on Ukrainian territory, as noted in a December 2024 McKenzie Intelligence Services (MIS) report commissioned by Greenpeace Ukraine.524
• Lack of appropriate personnel needed to run a nuclear facility. Qualification and staffing levels of current Zaporizhzhia plant personnel are uncertain.525 In addition, reportedly, Ukrainian staff has been subjected to violence, severe psychological pressure, and, in some cases, torture.526 Research undertaken by the Ukrainian NGO Truth Hounds documents cases of systematic detention, mistreatment, and torture of citizens associated with the plant, which it says, “likely constitute war crimes and crimes against humanity.”527
• Equipment degradation at Zaporizhzhia resulting from a combination of shortages of staff that carries out maintenance and inspection activities,528 onsite military activity, poor maintenance, and improper operation. During two winters, Russia kept one or two units in hot shutdown—a non-standard, resource-intensive mode that increased safety risks and contributed to equipment failures, including damage to steam generators, according to Energoatom.529 In April 2024, the last reactor was placed in cold shutdown.530 One notable case of infrastructure damage was the August 2024 fire in a cooling tower that caused serious harm, due to which, according to IAEA’s Rafael Grossi, the tower might have to be demolished.531

Since early 2023, Rosatom representatives have repeatedly declared intentions to restart reactors at Zaporizhzhia. Beyond the potential danger of operating any nuclear reactor in a war zone (see above), the further degraded conditions under military occupation, severely restricted access to reliable power and water supply as well as the ongoing hostilities around the plant, restarting units at the Zaporizhzhia site would bring risk levels to different dimensions as described in detail in a 2024 Bellona Foundation report.532 Such a decision could only be made at the highest political level in Russia. Meanwhile, technical preparations are underway; as of mid-2025, preparations in the form of grid-related activities have been observed in the occupied territories.533 Moreover, statements by the head of Rosatom indicate preparations for the construction of a floating pumping station with a capacity of 80,000 m³ per hour.534

It should be stressed that while the situation at Zaporizhzhia is exacerbated by the military occupation and the plant’s proximity to the frontlines, the other three operational nuclear power plants in Ukraine—comprising a total of nine reactors—are exposed to similarly serious fundamental safety and security risks. The potential of a dangerous event at an operating reactor is significantly higher than at a unit in cold shutdown.

Although these other plants have not come under direct attack during the war, they still face serious risks due to missile overflights and strikes on critical substations in the power grid. Such incidents have already forced reactors to reduce output, undergo emergency shutdowns, or disconnect entirely from the grid. Since autumn 2024, following an agreement with Ukrainian President Volodymyr Zelenskyy, the IAEA has also undertaken the task of inspecting key substations in Ukraine’s energy system that are critical for nuclear safety.535

In addition to the threats facing Ukrainian nuclear power plants, as mentioned above, ongoing hostilities and Russian attacks across Ukraine pose risks to other nuclear facilities, in particular:

• The Kharkiv research reactor has been subjected to repeated shelling leading to structural damage, loss of external power, and serious concerns about the security of nuclear materials and the safety of the facility.536
• In February 2025, a Russian Shahed drone struck the New Safe Confinement structure over Chornobyl’s Unit 4, creating a 15-square-meter hole in the roof and sparking fires that burned for over two weeks. Although radiation levels remained stable, the incident caused serious structural damage, prompting the IAEA to call for major repairs and raising international concern over nuclear facility security.537

Thus, while Russia is a long-standing and prominent member of the IAEA and a leading player in the international nuclear market through its state corporation Rosatom, it is at the same time—often with Rosatom’s direct involvement—engaged in unprecedented hostile actions in Ukraine that pose serious threats to the civilian nuclear sector. These include the seizure of nuclear power plants, attacks on critical energy infrastructure, and challenges to the non-proliferation regime through the capture of large quantities of foreign nuclear material.

United Kingdom Focus

The incoming government, since its election in the summer of 2024, has established new mechanisms and policy changes to deliver climate change and energy security objectives. This has included making available more government funds for constructing two reactors at Sizewell C and SMRs, as well as additional finance for nuclear fusion. This is despite the sector continuing to be plagued by the same old problems of cost overruns, delays in its new projects, and rising costs associated with the existing fleet.

As of mid-2025, the United Kingdom (U.K.) operates just nine reactors, the same as in the previous edition of the WNISR. The average fleet age is 38.1 years (see Figure 44). The last reactors to close were the two units at Hinkley Point B on 6 July 2022 (B-2) and 1 August 2022 (B-1), respectively. This followed the closure of the two reactors at Hunterston in 2021 and 2022, and two units at Dungeness officially closed in 2021 (last power generation in 2018, see Figure 43).

In total, 36 nuclear reactors have been closed, including all 26 Magnox, the first-generation gas-cooled reactors, two fast breeders, one small Steam-Generating Heavy Water Reactor (SGHWR), and seven Advanced Gas Reactors (AGRs). There is now 5.8 GW of nuclear capacity in operation, with 7.8 GW awaiting or undergoing decommissioning.

1. U.K. Reactor Startups and Closures

Sources: WNISR with IAEA-PRIS and EDF Energy, 2022–2025

Type of Reactors: AGR: Advanced Gas Reactors; FBR: Fast Breeder Reactor; PWR: Pressurized Water Reactor; SGHWR: Steam-Generating Heavy Water Reactor.

In 2024, nuclear power provided stable 38.6 TWh (40 TWh gross), contributing 14.2 percent of the total supply (up 0.3 percentage points), half the contribution of wind, and down from a maximum share of 28 percent in 1997.

Eight of the nine operating reactors are AGRs in pairs at Torness, Heysham (four reactors), Hartlepool, and one PWR at Sizewell. All operating AGRs were completed in the 1980s, while Sizewell B started operating in 1995 (see Figure 43).

Managing reactors as they age is a constant challenge of any technology design, and the AGRs are no exception. As has been commented on in previous editions of the WNISR, issues with the core’s graphite moderator bricks have raised continuous concerns.538

In January 2024, Electricité de France (EDF) announced that it was planning to extend the operating lifetimes of eight reactors and that a decision would be taken by the end of 2024.539 In December, EDF announced that they planned to operate Heysham-1 and Hartlepool until March 2027, while Heysham-2 and Torness AGR reactors could operate until March 2030 (see Table 11).540 EDF acknowledged that given the technical uncertainties of graphite aging the future generation from the AGRs “remains highly uncertain”.541 The other reactor, at Sizewell B, is scheduled to operate until 2035, but an investment decision is expected in 2025 to enable a 20-year lifetime extension.542 Although, EDF notes that they need greater cost certainty and confidence in the commercial case before making the decision.543 As of early 2025, to keep the reactors operating, EDF said it had invested £8 billion (US$10.8 billion) since 2009 and plans to make available £1.3 billion (US$1.8 billion) more between 2025 and 2027.544

1. Status of EDF’s U.K. AGR Nuclear Reactor Fleet (as of 1 July 2025)

Reactor	Net Capacity (MW)	Grid Connection	Closure/ Expected Closure
Dungeness B-1 Dungeness B-2	545 545	03/04/1983 29/12/1985	Closed June 2021 (Last power in 2018)
Hartlepool A-1 Hartlepool A-2	590 595	01/08/1983 31/10/1984	March 2027
Heysham A-1 Heysham A-2	485 575	09/07/1983 11/10/1984	March 2027
Heysham B-1 Heysham B-2	620 620	12/07/1988 11/11/1988	March 2030
Hinkley Point B-1 Hinkley Point B-2	485 480	30/10/1976 05/02/1976	Closed July/August 2022
Hunterston B-1 Hunterston B-2	490 495	06/02/1976 31/03/1977	Closed November 2021 Closed January 2022
Torness-1 Torness-2	595 605	25/05/1988 03/02/1989	March 2030

Sources: WNISR and EDF, 2025545

In April 2025 the Office for Nuclear Regulation (ONR) had moved Hartlepool into ‘enhanced regulatory attention’ to achieve improved performance in specific areas, such as “conventional health and safety, the number of site incidents and the production of nuclear safety cases.”546

1. Age Distribution of U.K. Nuclear Fleet

Sources: WNISR with IAEA-PRIS, 2025

U.K. Power Mix

The year 2024 broke several records in the electricity sector:

• Wind power reached the largest-ever share on the grid at any one time, providing over 68 percent of the country’s electricity on 18 December 2024.547
• Wind provided a total of 84 TWh of electricity and is now very close to natural gas – providing 86 TWh, as the leading provider of power. The contribution of coal in 2024 was just 1.9 TWh, with the final station closing mid-year—this is down from 100 TWh a decade ago. Nuclear provided 40 TWh.548
• CO2 emissions from the sector were also at a historic low, with the carbon intensity averaging just 125g CO2/kWh.549

While wind production growth has slowed down in recent years, additional large offshore wind capacity is expected to come online in 2025 and 2026.550

1. Electricity Generation by Source in the U.K. – The Coal Plunge

Source: U.K. Government, 2025551

The U.K. has set some of the world’s most ambitious greenhouse gas emission targets, committing to a 68 percent reduction from 1990 levels by 2030 and an 81 percent reduction by 2035.552

The Labour Government

On 4 July 2024, a Labour Government was elected for the first time in over 15 years. Under Prime Minister Sir Keir Starmer, the government made six key commitments, including one related to energy: the establishment of a new financial vehicle, ‘Great British Energy’, to help finance the energy transition.

Ed Miliband was re-appointed as the Secretary of State for Energy Security and Net Zero—a post held both in opposition and in the previous Labour Government, so he has considerable experience with the brief. He rapidly set out to reform the power sector and reconfirmed the election pledge to make the U.K. a “clean energy superpower”. This included a commitment to have a zero-carbon electricity system by 2030, which included plans to “double onshore wind, triple solar power and quadruple offshore wind.” To support these and other developments, the government is establishing a publicly-owned energy company—Great British Energy—which was capitalized by the government with £8.3 billion (US$11.2 billion).553 Although it later transpired that this would also fund nuclear projects.

In addition, the government lifted the de facto ban on onshore wind and changed grid connection rules so that projects ready to go were prioritized over those higher up the queue.554 In December 2024, the government published a Clean Power 2030 Action Plan, which, if fully implemented, would see the power sector generate 95 percent from “clean sources” in a “typical weather year”, with some 35 GW of unabated gas in reserve capacity.555

On nuclear, the plan stated

We are also committed to nuclear, including the lifetimes of existing nuclear projects where possible and the development of emerging low-carbon and renewable technologies that will play an important role beyond 2030.

The government further stated that

decisions on Sizewell C and the Great British Nuclear-led Small Modular Reactor programme will be taken at the Spending Review.

which took place in June 2025 (see following sections).556

Nuclear Newbuild

The U.K. has one power station with two reactors under construction at Hinkley Point C, and one project also with two units awaiting a Final Investment Decision (FID) at Sizewell C as of mid-2025.557 Both projects are based on the Franco-German European Pressurized Water Reactor (EPR) design.

Hinkley Point C

As has been documented in earlier editions of the WNISR, the construction of Hinkley Point C (HPC), instead of being the beginning of a bright new future for the nuclear industry, is the flagship project that exemplifies the scale of problems plaguing the U.K. nuclear program. It was initially estimated at £201618 billion (US$201624.3 billion) in 2016.558 At the time, Vincent de Rivaz, then chief executive of EDF, said, “We have the expertise, the supply chains and the teams ready to build Hinkley Point C safely, on time and on budget.”559 Since then the project’s costs have escalated dramatically to between £31 billion and £34 billion (US$47–52 billion), in 2015 values,560 in 2025 money this is £43–47 billion (~US$58–64 billion). This represents an increase of over 150 percent from the original budget, fundamentally undermining the economic case for new nuclear power. The rising construction cost is of concern to EDF, and it is seeking to find additional investors.561 In June 2025, EDF announced that it had signed an agreement with Apollo to provide an additional £4.5 billion (US$6 billion) in financing to assist with, among other things, the completion of HPC. The bonds would be issued in three tranches, starting with £1.5 billion (US$2 billion) in June 2025 and then an equal amount in the following two years.562

The project timeline has suffered equally significant setbacks. Scheduled at construction start for completion in 2025, the first unit is not expected to begin operations before 2029 at the earliest, with more realistic scenarios from EDF pushing this to 2031. The delays stem from multiple factors, including the complexity of civil engineering work, construction conditions during the COVID-19 restrictions, and the challenge of restarting nuclear construction in Britain after a 20-year hiatus.

In 2013, to prepare the project for the Final Investment Decision, the government introduced a price guarantee for 35 years, a Contract for Difference (CfD), depending on the number of units ultimately built. If only HPC were built, the strike price would be £92.5/MWh (US$2012146/MWh), but if Sizewell C was also built, the price for both would be £89.5/MWh (US$2012141/MWh). However, the figure was not fixed; the £92.5/MWh is to rise in line with the Consumer Price Index from 2013.563 Therefore, if the plant had operated in 2025, the price EDF would have received would have already increased to £132/MWh (US$2025178/MWh), and by 2030, will reach around £150/MWh (US$2025202/MWh). Given that the CfD is expected to operate for 35 years, EDF can expect prices of around £350/MWh (US$2025472/MWh) by 2065. In comparison, the latest round of offshore wind (AR6) in September 2024 awarded contracts with strike prices ranging from £201254–59/MWh (US$201285–93/MWh) with an outlier floating wind turbine project at £2012140/MWh (US$2012221/MWh).564 Furthermore, wholesale prices, which are currently predominantly set by the price of natural gas, are expected to fall as more renewable sources of electricity are brought online.565 According to the Financial Times, consultancy Aurora Energy Research estimates wholesale electricity prices would be in the region of ~£78/MWh (US$2025106/MWh) in the 2030s.566

Within the original agreement, if commercial operation starts after November 2029, the CfD is reduced by one year for every year of delay until 2033. This is the “longstop date”, after which the contract could be cancelled if the project is not completed. On 29 November 2022, the longstop date was extended from 1 November 2033 to 1 November 2036.567

The delay in completion has had an impact not only on EDF but also on the broader power market, which was expected to have 3.2 GW of capacity available. An assessment from early 2024 by Aurora Energy Research looked at the market implications of a three-year delay at Hinkley. It concluded that the wholesale price of electricity would rise by £20245/MWh (US$20246.7/MWh) in 2029, assuming it was replaced by gas and an increase of 6.2 Mt of CO2. However, they also concluded that the

cost of not paying for Hinkley C’s CfD is substantial, adding up to £1.62022 billion [US$20222 billion] over the three years we are considering here [2028–2030]. That offsets a bit less than half the cost of rising wholesale prices from the delay to Hinkley C.

The assessment concluded that

in short, the delay of Hinkley C is bad news for everyone, unless they happen to own an unabated gas plant or an interconnector, of course.568

Sizewell C

Initially, it was proposed that EDF and China General Nuclear Power Co. (CGN) would develop a follow-on to HPC, the Sizewell C or SZC project, a copy of HPC, with two 1.6-GW EPR units. Chinese investment was limited to 20 percent, leaving EDF with 80 percent of the company that would take the project to FID; neither party was obliged to take any share in the company that built, owned, and operated it. In 2022, EDF stated that it had planned to pre-finance the development of its share of the initial budget of up to £458 million (US$2022565 million), with no agreement to invest beyond that stage.569 On 24 June 2020, the U.K. Planning Inspectorate accepted the application for development consent received the previous month.570 In July 2022, the government gave its development consent to build Sizewell C.571

EDF was optimistic that it could reduce construction costs, and in 2020—less than two years into the official construction of HPC—it estimated that they would be £18 billion (US$202023 billion).572 However, it was also hoping to reduce the financing costs of Sizewell C by shifting from the CfD mechanism to the Regulated Asset Base (RAB) model. EDF has suggested that with a better financing model and no “first-of-a-kind costs” (FOAK)573, the strike price could “peel away” by £36/MWh (US$201256.9/MWh).574 In its 2020, planning documents, EDF confirmed construction cost estimates of “circa £20 billion” (US$202025.6 billion).575

Then in early 2025, the Financial Times reported that the cost of the project was expected to be around £40 billion (US$54 billion), “double the £20bn [US$25.6 billion] estimate given by developer EDF and the UK government for the project in 2020.”576 Meanwhile, the French state auditor had recommended that it not take a final investment decision in Sizewell until it has reduced its financial exposure to HPC.577

In March 2021, EDF’s financial report for 2020 said an FID was likely to be made in mid-2022 but used cautious language on the whole about the project, stating “to date, it is not clear whether the group will reach this target,”578 which it did not.

There are two companies involved in the development of the project. A development company that took the project to the FID—Sizewell C Limited—and the company that then owns the project. The ownership structure of these two companies does not need to be the same. The government has bit by bit bought into the project and invested in the development company and has said it will also take a significant share of ownership of Sizewell. The government’s past, current and expected financial support includes:

• In June 2022, the government announced that it had taken out the £100 million (US$2022123 million) option which would be converted into equity to take a 20 percent share in Sizewell C, should the project reach an FID.579
• In November 2022, the government made its Investment Decision and confirmed it was investing a further £679 million (US$2022837 million.580 In the summer of 2023, the government announced it invested a further £511 million (US$2023635 million) taking the total, at this stage, to £1.2 billion (US$20231.5 billion).581
• On 15 January 2024, the Sizewell C Development Consent Order was triggered, opening the path for construction start, despite no FID. Later in January, the government made available £1.3 billion (US$20241.7 billion)—bringing the total government investment to £2.5 billion (US$3.4 billion)—to enable infrastructure development work, such as roads and railways, to be undertaken prior to any FID.582
• In August 2024, the government announced the subsidy would be up to £5.5 billion (US$20247 billion)583 and allocated a further £1.2 billion (US$20241.5 billion) in September/December 2024584 then bringing the total to £3.7 billion (US$20244.7 billion).
• In April 2025, a further £2.7billion (US$3.7 billion) was announced, to be taken from the £5.5 billion (US$7.4 billion). In total by April 2025, £6.4 billion (US$8.7 billion) had been spent and or handed to the Sizewell C development company, leaving some £1.6 billion (US$2.2 billion) from the £5.5 billion potentially still to be allocated.

In the spending review the government announced a further £14.2 billion (US$19.2 billion) for Sizewell C between now and the financial year 2029–30.585 By the end of the decade, the government therefore intends to have allocated £17.9 billion (US$24.2 billion) for the construction of Sizewell C. Even if there is no delay, construction will still be ongoing at that time and additional government funding can be expected to be needed.

Funding was not the only issue to be resolved, and in March 2024, a government standalone company, Sizewell C Ltd, signed a deal with EDF to purchase the freehold of the land for the new power plant, followed in April 2024 by Framatome signing contracts “worth multi-billion euros” with Sizewell C Ltd for the delivery of the nuclear steam supply systems, the safety instrumentation and control systems, long-term supply of nuclear fuel, and maintenance services.586 The land ownership question was one of the unresolved issues that had held up ONR’s issuing of a site license, which was finally granted in May 2024.587

The 2025 Spending Review, also stated that the government announced that it was “ending decades of delay on critical energy projects by giving the green light to Sizewell C, a Small Modular Reactor, and Carbon Capture, Usage and Storage programmes.”588 The additional nuclear spending were: 589

• £2.5 billion (US$3.4 billion) for the development of SMRs, with Rolls-Royce selected as the preferred bidder (see Small Modular Reactors below).
• £2.5 billion for nuclear fusion, to be split between research centers in Oxfordshire, West Burton and Cumbria.590
• Less commented in the media was the £13.9 billion (US$18.8 billion) for the Nuclear Decommissioning Authority (NDA) to continue decommissioning and waste management activities at the existing sites.

Therefore, civil nuclear has received £33.1 billion (US$44.9 billion), over a quarter of the total economy-wide capital expenditure of £120 billion (US$162.6 billion) within the entire spending review.

It was expected that a French-U.K. nuclear summit would be held in July 2025, at which the Financial Investment Decision and funding from third parties would be announced. Rather the Financial Times, ran a story that suggested that the U.K. Government would increase its stake in the plant to 47.5 percent, with Brookfield Asset Management taking 25 percent, Centrica 15 percent and EDF further reducing its share to 12.5 percent.591 As of 1 July 2025, details were yet to be confirmed by the U.K. Government.592

Other Sites

Other sites have been proposed and developed to various degrees over the years. This includes Moorside in Cumbria being investigated at some point by Toshiba-Westinghouse as well as Hitachi-GE owned Wylfa Newydd on Anglesey and Oldbury on Severn in South Gloucestershire.

In March 2024, the previous government announced that it had, through Great British Nuclear, purchased the nuclear sites in Wylfa and Oldbury and that the Wylfa site was the preferred option for a third large-scale nuclear project after Hinkley and Sizewell.593

In February 2025, in a gung-ho press release titled, “Government Rips up Rules to Fire-up Nuclear Power”, the government announced it was going to change the planning rules for nuclear. In particular, to lift the restrictions on the construction of reactors in England and Wales that are currently limited to the siting of new reactors at the eight existing sites, to include SMRs in planning rules and to remove the expiry date of nuclear planning rules. The government hopes that this will change the rules to “back the builders”.594

Small Modular Reactors (SMRs)

The government has continued to promote SMRs as a means of meeting future energy security and decarbonization objectives. This has gone beyond talk, and the government has made significant funding pledges.

In November 2020, to support the development of a potential next generation of reactors, the government proposed to feed up to £385 million (~US$500 million) in an Advanced Nuclear Fund, with up to £215 million (US$2020276 million) going to Rolls-Royce’s SMR program.595 This, in November 2021, led to Rolls-Royce announcing that it had received £210 million (US$2021289 million) in government funding, £195 million (US$2021268 million) in private funds,596 and an additional £85 million (US$2021117 million) from the Qatar Investment Authority the following month.597

Rolls-Royce is developing a now 470-MW reactor—thus technically, it does not fall under the SMR definition with nominal capacities in the 30–300 MW range. In 2021, Rolls-Royce hoped its technology would complete the Generic Design Assessment (GDA) process with U.K. regulators around September 2024 and deliver first power in the early 2030s,598 but the company concluded Step 2 in July 2024,599—out of three steps expected to take four years in total—and had already revised its expectations as of 2023, aiming for the conclusion of the final phase in August 2026.600 Per latest announcements, the process is now to end in December 2026.601

Rolls-Royce has suggested that the nth-of-a-kind reactor (after ten have been built) would cost in the order of £1.8 billion (US$2.4 billion) per 470-MW unit602 and would generate power at £201940–75/MWh (US$201951–96/MWh)603. In evidence submitted in 2017, Rolls-Royce told the House of Lords, that 7 GW would “be of sufficient scale to provide a commercial return on investment from a U.K.-developed SMR, but it would not be sufficient to create a long-term, sustainable business for U.K. plc.” The House of Lords concluded: “Therefore, any SMR manufacturer would have to look to export markets to make a return on their investment.”604 In May 2024, the Polish government announced that it had decided in principle on ordering the Rolls-Royce SMRs, although no timetable was made clear at this time.605 In June, Vattenfall announced that it was on a shortlist of two for an SMR order in Sweden.606 In September 2024, the Czech Republic selected Rolls-Royce from a choice of 7 designs to be taken forward for further review with the objective of building reactors near Temelín in the 2030s.607 And in March 2025, CEZ Group became a shareholder in Rolls-Royce SMR, subject to terms and conditions to progressively acquire 20 percent of the company.608 The agreement is to supply up to 3 GW of electricity.609

Progress is also being made on the licensing of other reactor designs. In December 2023, the Office for Nuclear Regulation (ONR) began its GDA for a different reactor design from Holtec International,610 and then in January 2024 for GE’s BWRX-300 design.611 In December 2024, the ONR, the Environment Agency, and Natural Resources Wales announced that the BWRX-300 reactor, being designed by GE-Hitachi, had progressed to Step 2 of the GDA.612

As mentioned above, in June 2025, the Spending Review, allocated £2.5 billion (US$3.4 billion) for SMRs and to Rolls-Royce in particular, to identify a site later this year and connect projects to the grid in the mid-2030s. The funding will be for three SMR units. Simon Bowen, the interim chair of Great British Energy-Nuclear told the Financial Times that he would have liked to back more than one developer, but there was not any money.613 Even this money was taken from Great British Energy, which was, when established in 2023 by the previous administration, not expected to fund nuclear projects.614

Over the past 12 months, the government and private investors have considerably increased funding available for nuclear power in the U.K., and there has been agreement for the extension of the operating lives of existing reactors, so by some measures, it has been a successful year for the industry. However, despite this, by the end of the decade, nuclear contribution to the power sector is expected to decrease as the power sector is expected to be close to fully decarbonized, mainly due to the contribution of renewable energies.

United States Focus

Overview

With 94 commercial reactors operational, the United States still has the largest nuclear fleet in the world. Nuclear energy generation in 2024 increased slightly—+0.9 percent—to 782 TWh. The nuclear sector’s share of utility-scale electricity generation decreased to 18.2 percent, matching the historic low (since the 1980s) of 2022. When generation supplied by small-scale solar PV is included, nuclear provided only 17.8 percent of total generation, 0.2 percentage points less than in 2022.615

Despite two new reactor startups between April 2023 and March 2024, the U.S. fleet continues to age, with a mid-2025 average of 43.7 years, making it amongst the oldest in the world: 57 units have operated for 41 years or more (of which 20 for more than 51 years) and all but four for 31 years or more (see Figure 46).

1. Age Distribution of U.S. Nuclear Fleet

Sources: WNISR, with IAEA-PRIS, 2025

Since the March 2024 grid connection of the second Westinghouse AP-1000 reactor at Plant Vogtle—Unit 4—there are no new reactors under construction in the U.S. Construction permits for two new commercial nuclear power plants have been submitted to the Nuclear Regulatory Commission (NRC) since March 2024:

• The 340-MW TerraPower Natrium reactor project proposed to be built in the State of Wyoming: TerraPower submitted the Natrium application in March 2024, and the NRC reports that the review is expected to be completed in 2026.
• The 320-MW Long Mott project to be built in Texas. The Long Mott permit application is for four 80-MW small modular reactors and was submitted in March 2025. The reactors would provide power and process heat to an existing Dow chemical facility.

The U.S. Department of Energy (DOE) is sponsoring both projects through over US$1 billion awards under the department’s Advanced Reactor Demonstration Program.

Several factors have led to a perception that new market opportunities are opening up for nuclear generation in the U.S.: the availability of federal financing and subsidies; expressions of political and economic support from state government officials; and projections of sharply increasing electricity demand. This has resulted in a few proposals to restart reactors after permanent retirement. Licensing reviews have progressed in the first of these cases, to restart the Palisades reactor from its retirement in May 2022. The restart plan and its total cost were somewhat clouded by the findings of a mid-2024 inspection of the reactor’s steam generators, which identified around 1,000 cracked tubes that must be plugged or repaired before restart.616 In February 2025, Holtec applied for a license amendment to authorize an alternative to plugging the damaged tubes by installing stainless steel sleeves.617 The NRC reports that it could issue the license amendment as soon as September 2025,618 which could enable restart by the end of the year, according to Holtec.619 In addition, the owners of the retired Duane Arnold-1 (2020) and Three Mile Island-1 (2019) reactors intend to submit restart applications to the NRC.

The submission rate of applications for license extensions declined between mid-2024 and mid-2025. The proposal to extend operation of the Diablo Canyon-1 and -2 reactors for five years continues to progress, with the ongoing review of Pacific Gas and Electric Company’s (PG&E) license renewal application. No applications for Initial License Renewal (ILR) were submitted in the period, and only two applications (for a total of three units) for Subsequent License Renewal (SLR)620 from 60 to 80 years were submitted.621

However, owners of 30 more reactors have notified the NRC they intend to file applications for license renewals in the coming years: one reactor for initial license renewal and 29 for subsequent license renewal. Between applications that are approved, currently under review, or pending submission, owners have decided to pursue SLRs for 51 of the 91 currently operating reactors built before the year 2000. In addition, owners are also planning to seek SLRs for three reactors that are currently proposed to restart from permanent shutdown: Duane Arnold-1, Palisades-1, and Three Mile Island-1.622

Subsidies and Financing for Nuclear Power

Following the January 2025 inauguration of President Donald Trump, there was renewed uncertainty about the availability of federal subsidies for the nuclear industry, clouding the potential for both new reactor construction and retirements of existing reactors. The principal subsidies of interest to the nuclear energy industry are:

• the Nuclear Power Production Tax Credit (Nuclear PTC), which provides tax credit subsidies to nuclear reactors built before 16 August 2022;
• the Clean Electricity Production Tax Credit (CE PTC) and the Clean Electricity Investment Credit (CE ITC), which provide tax credits to new “clean” generation sources, including new reactors;
• the Clean Hydrogen Production Tax Credit (H2 PTC), tax credits for producing hydrogen through processes with low/zero greenhouse gas (GHG) emissions.

2024 was the first tax year for which the Nuclear PTC subsidy can be claimed, and it has perhaps the greatest implications for the industry. Under the language of the IRA (Inflation Reduction Act), the value of the subsidy is reduced for reactors that earn revenues over US$25/MWh during the tax year, zeroing out at US$43.75/MWh.623 According to Constellation, the largest U.S. nuclear power plant owner/operator, the subsidy essentially underwrites the continued operation of existing reactors by creating a type of revenue floor through 2032. It said that

The nuclear production tax credit (PTC) in the IRA provides a stable foundation for consistent and growing earnings that will allow Constellation to continue investing in growth opportunities, including by adding clean energy generation to its fleet through uprates, repowering wind assets, license extensions and asset acquisitions while also returning capital to shareholders. The PTC provides revenue visibility and also preserves Constellation’s ability to capture upside from tightening power market conditions.624

While it is unclear what the new government may do to roll back incentives for renewable energy, there are indications that the administration may not move to curtail programs that support nuclear power.625 One of the earliest signals was the Trump administration’s affirmation of financing and subsidies to support a proposed restart of the retired Palisades-1 reactor. In March 2025, DOE disbursed the second installment of a US$1.5 billion loan guarantee,626 and the U.S. Department of Agriculture (USDA) affirmed the award of US$1.3 billion to rural energy cooperatives in the states of Michigan and Indiana for the purchase of electricity from Palisades.627 Together with US$300 million in grants from the State of Michigan,628 Palisades’ restart is to receive more than US$3 billion in public sector support.

The USDA grant is effectively a subsidy to Holtec, helping the cooperatives cover the excessive cost of power from Palisades. The reactor closed in 2022, when a 15-year Power Purchase Agreement (PPA) between Entergy and the utility Consumers Energy (CE) expired.629 CE was unwilling to continue the PPA at what had proven to be significantly greater than wholesale market prices,630 averaging US$68/MWh in the final year,631 while annual average market electricity prices in the Midcontinent ISO had not exceeded US$40/MWh over the previous decade.632 Entergy was uninterested in continuing to operate Palisades in a competitive market environment, and in 2017 announced its intent to retire the reactor in Spring 2022.633 In Summer 2024, inspections necessary for the restart effort revealed extensive corrosion of steam generator tubes, requiring nearly 1,000 tubes to be plugged.634 Reportedly, according to a NRC representative speaking at a public meeting reviewing Holtec’s plans, the repairs could delay the restart schedule.

States and the federal government continue to provide large volumes of financial support for both new and existing reactors. However, states that adopted nuclear subsidy programs prior to the IRA have taken steps to mitigate the costs of those programs. Via a complex formula in the statute, the Nuclear PTC is limited by the amount of other revenue the reactor receives, zeroing out at earnings of US$43.75/MWh; however, the IRA permits reactor owners that receive state government subsidies to claim up to the full value of the PTC, if the company uses the funds to defray the cost of the state’s subsidy. The subsidies in both Illinois635 and New Jersey636 are capped at values of about US$10/MWh, less than the maximum amount of the Nuclear PTC. The recently adjusted amount of New York’s nuclear subsidy is US$14.70/MWh, less than the maximum value of the Nuclear PTC.637 As a result, the net cost of each of these state’s subsidy programs is expected to reach US$0 in 2025.

In addition, Massachusetts enacted legislation in 2024 that authorizes the state’s utilities to enter into long-term contracts to purchase electricity from nuclear power plants in the neighboring states of Connecticut and New Hampshire.638 The legislation stems from a negotiation initiated by the governor of Connecticut to get Massachusetts to share the cost of his state’s power purchase contracts with the Millstone Nuclear Power Plant, the costs of which have become unpopular with Connecticut consumers.639

The U.S. Congress has enacted other measures to promote the expansion of nuclear energy, the development and licensing of new reactor designs, and exports of nuclear reactors. President Biden signed the Prohibition on Russian Uranium Act in May 2024, which contains provisions that unlock US$2.7 billion for domestic production and/or procurement of enriched uranium from non-Russian sources, including High-Assay Low-Enriched Uranium (HALEU).640 The HALEU issue raised eyebrows within the U.S. non-proliferation community.641 The legislation offers the industry waivers through 2028, if reactor owners continue to rely on Russia for fuel. In another budgetary measure enacted in March 2024, Congress renewed and extended the Price-Anderson Act nuclear liability statute for another forty years.642 Under a previous extension of Price-Anderson enacted in 2005, all existing reactors enjoy its protections in perpetuity with no further legislation required, but industry desired the new extension to ensure all reactors built after 2025 will be covered, as well.

Other Support Measures for the Nuclear Industry

In July 2024, President Biden enacted a bill, S. 870,643 which includes the Accelerating Deployment of Versatile, Advanced Nuclear for Clean Energy (ADVANCE) Act of 2024, an omnibus nuclear energy bill.644 The ADVANCE Act does not contain any direct subsidies for nuclear energy, but it includes several measures intended to promote nuclear energy and relax regulations on nuclear safety and licensing. It supports a U.S. Government agenda to compete with Russia and China for reactor exports, through tasking the NRC with creating the “International Nuclear Export and Innovation Branch” (within the Office of International Programs) to collaborate with domestic agencies involved in nuclear export deals, interact with international governance bodies, and assist other countries in establishing their own regulatory systems. It also relaxes the Section 810 procedures for authorizing nuclear technology exports that were established to ensure compliance with the Nuclear Nonproliferation Treaty; relaxation of Section 810 could expedite export deals but raises concerns about the potential for exporting technologies like uranium enrichment that could enable the proliferation of nuclear weapons.

The ADVANCE Act does away with the prohibition on foreign ownership of nuclear power plants in the U.S. for OECD (Organization for Economic Cooperation and Development) members and India, while prohibiting NRC from approving imports of enriched uranium from Russia and China without prior authorization by the Secretary of Energy and Secretary of State. The new statute also requires the NRC to modify its mission statement to include language stating that:

licensing and regulation of the civilian use of radioactive materials and nuclear energy be conducted in a manner that is efficient and does not unnecessarily limit—

(1) the civilian use of radioactive materials and deployment of nuclear energy; or

(2) the benefits of civilian use of radioactive materials and nuclear energy technology to society.

The act requires NRC to hire more employees to review license applications, expedite and relax its licensing procedures, and to exclude more topics from consideration in environmental reviews by expanding NRC’s practice of issuing “categorical exclusions.”

Extended Reactor Licenses

Under the Atomic Energy Act (AEA) and NRC regulations, the NRC issues initial operating licenses for commercial power reactors for 40 years. Regulations provide for license extensions in increments of up to 20 additional years. Due to the advanced age of the U.S. reactor fleet, reactor owners have begun applying for second license extensions, to permit operation out to as much as 80 years.

As shown on Figure 47, as of mid-2025, 86 of the 94 operating U.S. units (91.5 percent) had already received 20-year Initial License Renewals (ILRs), which permit reactor operation beyond 40 and up to 60 years. The NRC previously approved ILRs of three currently retired reactors that are the subject of proposed restarts (Duane Arnold-1, Palisades-1, and Three Mile Island-1). Owners have filed ILR applications for four of the eight operating reactors which are not (yet) recipients of license renewals.645 The other four reactors have not been operating long enough to qualify for ILRs: Watts Bar-1 and -2, which began operation in 1996 and 2016, respectively; and Vogtle-3 and -4, which entered commercial operation within the last 24 months. TVA has notified NRC of its intention to apply for an ILR for Watts Bar-1 by the end of 2026.646

1. Status of License Renewal Applications in the U.S.

Sources: compiled by WNISR with NRC, 2025

Notes: ILR: Initial License Renewal; SLR: Subsequent License Renewal.

* Subsequent License Renewal applications expected for three closed reactors: 2025 (Duane Arnolds), 2026 (Palisades), 2029 (TMI-1).

As of mid-2025, the NRC had approved Subsequent License Renewals (SLR) for thirteen reactors, which would permit their operation from 60 to 80 years. Applications for a further twelve reactors are under review,647 and owners have notified of their intentions to submit applications for a further 29 reactors between 2025 and 2034; of those pending applications, three are for currently retired reactors subject to proposed restarts. (See Figure 47) Thus, owners are planning towards or have received second license extensions for 54 percent of currently operating reactors.648 It is unclear how owners’ plans to submit SLR applications may change if the Nuclear PTC subsidy program created by the Inflation Reduction Act is not extended past its current expiration date of 2032.

Reactor Closures and Proposed Restarts

The average age of the seven reactors closed in the U.S. over the five-year period 2018–2022—no unit was closed since—was 47.1 years (see Figure 48), significantly below their licensed lifetimes of 60 years.

1. Evolution of Average Reactor Closure Age in the U.S.

Sources: WNISR with IAEA-PRIS, 2025

The retirement of Palisades in May 2022 marked the thirteenth closure in ten years, starting with the retirements of four reactors in the first half of 2013. With the effort to extend the operation of Diablo Canyon-1 and -2, there are no further anticipated closures before 2029, when the operating licenses of the oldest reactors in operation expire (Nine Mile Point-1 and Ginna-1). Constellation has notified the NRC of its intent to seek SLRs for Nine Mile Point-1 and Ginna-1,649 but the company has already delayed submittal of those applications,650 likely due to uncertainty regarding the ongoing availability of subsidies: New York State’s ZEC program expires in March 2029, and the federal Nuclear PTC expires in 2032, just three years into the reactors’ SLR periods.

Restarts of Retired Reactors

The effort by Holtec and Michigan officials to bring Palisades-1 out of retirement would make it the first reactor in the U.S.—and only the second in the world after the restart of Armenian-2 (also Metsamor-2) in Armenia, several years after its closure in 1989—to return to operation after entering decommissioning. In recent months, the owners of two other reactor owners have notified the NRC that they plan to seek restarts: NextEra to pursue bringing Duane Arnold-1 back into service, after its retirement in 2020651; and Constellation plans to restart Three Mile Island-1.652 The latter is in consequence of an agreement with Microsoft to sell power from the reactor under a forthcoming 20-year Power Purchase Agreement (PPA). As part of the restart project, Constellation proposes to rebrand Three Mile Island-1 as the Crane Clean Energy Center, named after the deceased CEO of Exelon under whose tenure the reactor was retired in 2019. In both cases, the restart plans surfaced in connection with anticipated growth in electricity demand and the development of energy-intensive hyperscale data centers and artificial intelligence technology.653

The proposed restart of Palisades-1 could establish regulatory pathways and test the practical feasibility of the practice for the rest of the industry. Holtec faces significant obstacles to doing so:654 it has no experience building or operating nuclear reactors, its core businesses are in developing and manufacturing irradiated fuel storage systems and, since 2018, in managing the decommissioning of reactors. Palisades is known to have a long list of maintenance needs, such as the control rod drive seals, which would require significant expense for a new owner to take on.655 As reported above, inspections in Summer 2024 discovered extensive corrosion of steam generator tubes, requiring maintenance that could delay Holtec’s proposed restart schedule.

Because the restart of a reactor that has entered decommissioning is unprecedented in the U.S., the NRC licensing process may take longer than expected and be subject to interventions and legal challenges. Holtec has submitted nine applications for license amendments, technical specification changes, and other approvals necessary to resume operations.656 NRC projects that these reviews could be completed before September 2025, subject to completion of repairs and other contingencies.657

Figure 49 provides an overview of reactors retired between 2009 and 2022 (some of which are now subject to reversals) or previously considered for “early closure”.

1. U.S. Early Reactor Retirements and Some Reversals

Sources: Various, compiled by WNISR, 2025

Notes: Eight of the notices (Letters of Intent) provided to the NRC do not specify for which reactors the owners intend to submit applications and likely involves Fitzpatrick, Davis Besse-1 & -2, and possibly Byron-1 & -2; see above.

* Crystal River: No production after 2009 (WNISR considers it closed as of this date). Official closure announced in 2013. Renewal application submitted in 2008, withdrawn in 2013; see U.S. NRC, “Crystal River – License Renewal Application”, Updated 9 December 2016, see https://www.nrc.gov/reactors/operating/licensing/renewal/applications/crystal-river.html, accessed 8 September 2020.

** Potential restart from early closure. Future Subsequent License Renewal applications expected in 2025 (Duane Arnolds), 2026 (Palisades), 2029 (TMI-1).

*** Early closure reversed following access to new subsidies.

**** Initial License Renewal application canceled or withdrawn in 2018. New applications submitted in 2023, currently under review.

New Reactors: Proposals, Planning, and Policy Developments

After a 20-year pause, three reactors were connected to the grid between 2016 and 2024 (see Figure 50). Since the completion of Vogtle Unit 3 and 4, there are no new commercial reactors under construction in the U.S. for the first time since construction of Watts Bar-2 was reactivated in 2007 (see Figure 51).

1. Reactor Startups and Closures in the U.S.

Sources: WNISR with IAEA-PRIS, 2025

NRC is currently reviewing three construction permit applications. One reactor design received Standard Design Approval (SDA) in May 2025.658

• TerraPower’s Natrium project in Kemmerer, Wyoming, to build a 345-MW sodium-cooled fast neutron reactor based on General Electric’s PRISM design. The facility design includes a molten salt thermal storage loop capable of generating an additional 150-MW for up to six hours. TerraPower submitted the Construction Permit Application (CPA) in March 2024, and NRC expects the review to be completed in 2026.659 The company broke ground on a “test and fill facility” for the reactor’s sodium coolant/moderator in June 2024.660
• Dow Chemical’s Long Mott Generating Station (LMGS), a 320-MW plant comprising four X-energy Xe-100 high-temperature gas-cooled, graphite-moderated reactors. LMGS would be located at Dow’s UCC Seadrift petrochemical plant in Texas. The reactors would have a nominal electricity generating capacity of 320 MW and could produce up to 800 MWt of steam process heat. Dow submitted the CPA in March 2025.661 It projects a five-year construction schedule, with completion in 2033.662
• TVA’s Clinch River SMR project, to construct a 300-MW GE-Hitachi BWRX-300 reactor. In 2019, TVA received an Early Site Permit (ESP) to construct a then-unspecified SMR at the Clinch River Site, where U.S. DOE canceled construction of a demonstration breeder reactor in 1983. TVA completed submission of the CPA in May 2025.663
• Nuscale’s US-460, a standardized 462-MW facility comprised of six 77-MW light-water SMRs, received a Standard Design Approval (SDA) from the NRC in May 2025.664 In November 2023, due to cost escalation, NuScale canceled its Carbon-Free Power Project that would have demonstrated the US-460 design. The company continued to pursue SDA for the US-460, but it has no proposed projects or procurement deals to build it. SDA does not confer the same degree of technical review and approval as Standard Design Certification (SDC), as it “does not prevent issues resolved by the design review process from being reconsidered during a rulemaking for a design certification or during hearings associated with a construction permit or combined license application”; rather, NRC characterizes the SDA as providing “incremental progress towards the licensing or certification” of a reactor design.665

The only nuclear reactor construction projects currently underway are three prototypes of commercial reactors and the U.S. Department of Defense’s Project Pele, a 5-MW mobile microreactor demonstration project from BWXT Advanced Technologies.666 The prototype reactors are:

• In May 2025, Kairos began nuclear construction of its Hermes non-power test reactor (35 MWt), at Oak Ridge National Laboratory in Tennessee, a molten-fluoride-salt design;667
• In November 2024, Kairos received a construction permit for the two-unit Hermes 2 non-power test reactors (2 x 35 MWt), at Oak Ridge National Laboratory,668 but, as of mid-2025, has not announced the start of nuclear construction;
• In September 2024, NRC issued the construction permit for the Natura molten salt research reactor at Abilene Christian University (1 MWt).669 The university announced the opening of the research laboratory where the reactor will be housed in 2023, but, as of mid-2025, has not announced the start of nuclear construction. The research reactor is intended to demonstrate the feasibility of the molten salt technology, based on which Natura has reportedly struck agreements to develop commercial power reactors at Texas Tech University and Texas A&M University’s proposed RELLIS facility670—the latter, a planned test-bed for various new reactor designs for which the college is seeking an Early Site Permit at its Bryan, TX satellite campus.671

Six of the seven projects above have benefited from significant, direct federal government sponsorship. DOE’s Advanced Reactor Demonstration Program has awarded X-energy, TerraPower, and Kairos a combined US$2.8 billion in matching grants to construct their reactors; DOE is providing land to Kairos at Oak Ridge National Lab672 and to TVA at the Clinch River Site, and offered land for NuScale’s Carbon-Free Power Project until it was canceled in 2023;673 TVA itself is a federal government corporation with independent ratemaking authority, under which it sells power to municipal utilities and rural electric cooperatives across seven states; and Project Pele is funded by the Department of Defense, with the intent to develop mobile microreactor technologies that could also be deployed commercially.674 Furthermore, the Department of Defense has conducted a procurement for a microreactor to power the Eielson Air Force Base in Alaska,675 and the U.S. Army’s Corps of Engineers has recommended procurement of an SMR to power the Fort Drum Army Base in New York.676

The official schedules for two commercial reactor demonstration projects have slipped beyond 2030, TerraPower’s Natrium and X-energy’s Xe-100 reactor designs, funded by the Infrastructure Investment and Jobs Act (IIJA). The projects had been selected in 2020 as the flagship projects of DOE’s Advanced Reactor Demonstration Program (ARDP), with a goal of bringing reactors online in five to seven years.677 The ARDP is also supporting development of eight other power reactor designs, with goals for deployment of demonstration reactors in the early- to mid-2030s, at the soonest. The Department of Defense is sponsoring development of three microreactors, with projected operation dates in or around 2027.678

TerraPower submitted its application for a construction permit at its Kemmerer site, in Wyoming, to the NRC in March 2024. According to the agency, as of December 2024, its review of the application was less than half complete, with a projected completion date of December 2026.679 While the ARDP originally expected both the Natrium and Xe-100 reactors to be in operation in 2027, TerraPower now projects Natrium will begin commercial operation in 2031.680 TerraPower frequently states that the projected cost of the Natrium project is US$4 billion;681 however, in a July 2024 interview, the company’s co-founder, former Microsoft CEO Bill Gates, stated that its full development cost could be “close to US$10 billion”.682 With 345 MW of nuclear generation capacity, the cost would be US$12,000–29,000/kW. At the low end, the pre-construction estimate is nearly double that of the Vogtle 3&4 reactors at a similar stage, prior to construction (US$6,000/kW); at the high end, it is substantially greater than the final, all-in cost of the Vogtle reactors (US$16,000/kW).

In the past 45 years, there were only two construction starts of commercial reactors in the U.S. that were completed, at the Vogtle site, while construction of two more started but was later abandoned, at the V.C. Summer site (see Figure 51).

1. Seventy Years of Nuclear Reactor Constructions in the U.S.

Sources: WNISR with IAEA-PRIS, various dates

Note: Construction-start years refer to dates initially indicated in IAEA-PRIS database, subject to later modifications.

The X-energy demonstration project is being undertaken by a subsidiary of Dow Chemical, to be constructed at Dow’s UCC Seadrift chemical production facility in Texas. The four-reactor, high-temperature gas-cooled SMR plant would have a gross generating capacity of 320 MW, though it would also provide process heat (in the form of steam) to support UCC Seadrift’s production processes. Dow submitted a construction permit application for the reactors, to be titled Long Mott Generating Station, in March 2025. The application expresses hope that the NRC will issue the permit in 2027, that nuclear construction will begin in early 2028, and that the plant will begin operation in 2033.683

Many Projects in Early Stages—Modest Private Sector Investment

Private sector investment in new reactors remains limited primarily to research and development and venture capital, rather than commercial deployment. As a result, reactor design startup companies that have little if any business income continue to proliferate in the U.S., with ten reactor concepts added to the NRC’s pre-application engagement docket between 8 February684 and 9 June 2025.685 However, only one established commercial nuclear generation company (Tennessee Valley Authority, or TVA) has taken concrete steps toward building a new reactor: TVA’s selection of the GE-Hitachi BWRX-300 for its Clinch River site,686 for which TVA received an Early Site Permit (ESP) to build a then-unspecified SMR in 2019.687 Even that decision is tentative. In April and May 2025, TVA submitted a construction permit application for the reactor in two parts: the environmental report in April688; and the preliminary safety analysis report in May. The filing from 28 April 2025 states:

As communicated previously, the TVA Board has not yet authorized the deployment of a SMR at the CRN [Clinch River Nuclear] Site. TVA’s submittal of the CPA is an important step to de-risk the licensing aspect of a potential, future SMR deployment. Any decisions about deployment will be subject to support, risk sharing, required internal and external approvals, and completion of necessary environmental and permitting reviews.689

Several nuclear utilities have expressed interest in pursuing possible new reactor projects, for instance, through applying for ESPs or including new reactors in public utility commission planning dockets:

• Appalachian Power, a subsidiary of American Electric Power (AEP), initiated pre-application activities with the NRC in March 2025 for an ESP for an SMR690 at its Joshua Falls substation site;691
• AEP’s Indiana & Michigan Power subsidiary has included SMRs in its 2024 Integrated Resource Plan (IRP) filed with the Indiana Utility Regulatory Commission (IURC), and indicated that to “partially offset the cost of the early site permit process” it is seeking a grant from the DOE;692
• Constellation is seeking DOE funding to apply for an ESP at its Nine Mile Point Nuclear Power Plant, supported by the New York State Energy Research and Development Authority which pledged a matching share of the application cost;693
• Duke Energy has initiated pre-application activities with the NRC for an ESP at its Belews Creek coal power plant site,694 and the company included SMRs in the most recent IRP it submitted to the IURC;695
• Energy Northwest has initiated pre-application engagement with the NRC for a CPA to build up to 12 X-energy Xe-100 SMRs on the DOE’s Hanford Reservation, adjacent to Columbia Generating Station site, Washington,696 for which Amazon has reportedly committed US$334 million to fund a “multiyear feasibility study”;697
• PacifiCorp has included development of new reactors in IRPs its utilities have submitted to public utility commissions in Utah698 and in Wyoming (TerraPower’s Natrium project).699

None of these activities represent a formal investment decision. Even less, some seem contingent on further public financing. For instance, Constellation intends to fund its ESP application for Nine Mile Point entirely with government funding: 50 percent from the New York State Energy Research and Development Authority and 50 percent from a DOE grant program announced in late 2024.700

The majority of federal and state government investments are going toward prolonging the operation of existing reactors rather than into new construction. For instance, while existing nuclear generation companies have not ordered any new reactors and, with the exception of TVA’s Clinch River CPA, have not named specific plans to do so, since April 2024, six companies have announced plans to implement power uprates of 16–19 reactors, which would total more than 1200 MW in increased generating capacity:

• Constellation has stated it will implement power uprates of Braidwood-1 and -2 (135 MW),701 Byron-1 and -2 (135 MW),702 and Calvert Cliffs-1 and -2 (175 MW);703
• DTE Energy is considering to uprate Fermi-2 (150 MW);704
• Energy Northwest is planning an uprate of Columbia Generating Station (186 MW);705
• Entergy intends to implement up to 275 MW in uprates across its five-reactor fleet, starting with a 40 MW uprate of Waterford-3;706
• PSEG intends to pursue uprates of Salem-1 and -2 (168 MW),707 which it co-owns with Constellation; and
• Southern Co. announced that it will seek power uprates of Hatch-1 and -2 (58 MW) and Vogtle-1 and -2 (54 MW).708

Together with the industry’s plans to seek ILR and SLR operating license extensions, these indicators suggest the U.S. industry is focused on treading water with existing reactors rather than on significant amounts of new reactor construction within the next decade.

Nevertheless, the large amount of financial support in the Inflation Reduction Act (IRA) and Infrastructure Investment and Jobs Act (IIJA)—amplified by media coverage of deals between major Information Technology (IT) corporations and mostly startup reactor design companies—has correlated with political interest in pursuing new reactors. Since the IIJA was enacted in 2021, several states have enacted legislation and initiated programs to promote nuclear energy, and several utilities have initiated feasibility studies or included deployment of new reactors in their IRPs. The federal structure of U.S. governance provides states a significant role in energy infrastructure development—and a primary role where many investment decisions are concerned. This dynamic was evident in previous periods of nuclear construction, in which states’ enactment of Construction Work In Progress (CWIP) financing policies played a foundational role enabling utilities to proceed with nuclear projects. In the U.S.’s latest limited newbuild period, state CWIP policies were evidently more significant than the forgivable loans available through the Loan Guarantee program (see United States Focus in WNISR2023). Construction of the Summer-2 & -3, and Vogtle-3 and -4 projects both proceeded with CWIP financing, whereas utilities in South Carolina opted to move forward without seeking loan guarantees for the Summer reactors.

State policy developments may be bellwethers and enablers of future investment trends, and trends evidence efforts to lay groundwork for reactor construction. In some cases, states have gone so far as to authorize CWIP or similar financing mechanisms; in others, states are making investments that entail less risk exposure but are aimed at building workforce and/or industrial infrastructure that would facilitate reactor construction. Since 2023, six states have established programs or offices to promote development of new reactors and/or related industrial infrastructure:

• Kentucky: The state legislature overturned Governor Andy Beshear’s veto in 2024 to create the Kentucky Nuclear Energy Development Authority (KNEDA).709 The state also created a US$8 million Nuclear Energy Development Grant Program for FY2025–26 to be administered by KNEDA.710
• New York: Governor Kathy Hochul ordered the New York State Energy Research and Development Authority (NYSERDA) to commission a Draft Blueprint on Advanced Nuclear Energy.711 Following the publication of the final Blueprint in January 2025, Hochul announced the state’s support for Constellation’s ESP application for Nine Mile Point and further directed NYSERDA to create a Master Plan for Responsible Advanced Nuclear Development, which is expected in late 2026.712
• Ohio: In 2023, Ohio enacted legislation establishing the Ohio Nuclear Development Authority,713 with the stated purpose of making the state “a leader in the development and construction of new-type advanced-nuclear-research reactors; a national and global leader in the commercial production of isotopes and research; a leader in the research and development of high-level-nuclear-waste reduction and storage technology.”714
• Tennessee: Governor Bill Lee empaneled a Tennessee Nuclear Energy Advisory Council in 2023,715 which returned its Final Report and Recommendations in October 2024.716
• Texas: Texas Governor Greg Abbott directed the formation of an Advanced Nuclear Reactor Working Group in August 2023,717 which submitted a report in November 2024 with seven legislative recommendations to “bring ANR [Advanced Nuclear Reactor] projects to Texas.”718 Through legislation enacted in 2025, the state followed up the working group’s recommendations by establishing the Texas Advanced Nuclear Energy Office.719 In addition, in February 2025, Texas A&M University announced agreements with four reactor design companies to construct demonstration reactors at the RELLIS campus: Aalo Atomics, Kairos Power, Natura Resources, and Terrestrial Energy.720
• Virginia: In 2023, state legislation established a Nuclear Innovation Hub within the Virginia Power Innovation Program and created a Nuclear Education Grant Fund.721

In addition, through the National Association of State Energy Officials (NASEO), ten states have formed, in early 2025, a collaboration “to reduce the cost of advanced nuclear projects,” the Advanced Nuclear First Movers Initiative (ANMFI). Co-chaired by Indiana, Kentucky, New York, Tennessee, and Wyoming, the ANMFI also says it will seek to expedite federal licensing of projects and form public-private partnerships.722

Eight states enacted policy changes or initiated planning processes to promote nuclear energy since July 2024. The measures range from authorizing CWIP financing and providing direct subsidies, to classifying nuclear for clean energy incentives and conducting feasibility and planning studies:

• Arkansas: The legislature enacted SB 307, authorizing utilities to use CWIP-ratemaking to fund construction of nuclear reactors (as “strategic investments”).723 Another new law, HB 1572, directs the Arkansas Department of Energy and Environment to commission a “technical feasibility study on implementing nuclear energy generation.”724
• Colorado: The legislature enacted a bill (HB25-1040)725 classifying nuclear power as a “clean energy resource” toward the state’s energy policies and 2050 resource targets.726
• Indiana: Legislation enacted categorizes nuclear power as a “clean energy” or “green energy” resource (SB 178)727; creates a pilot program for SMRs under the Indiana Utility Regulatory Commission, with ratepayer financing (SB 423)728; amends the CWIP financing statute enacted in 2024 to authorize recovery of other project development costs, such as an Early Site Permit application (SB 424)729; and creates a subsidy for manufacturers of SMRs in Indiana, through a Small Modular Nuclear Reactor Manufacturing Expense Tax Credit (HB 1007).730
• Kentucky: A concurrent resolution introduced in the House (HCR 22) declares nuclear energy “a clean and dispatchable means of providing baseload electricity” and implies that it should be prioritized in state energy plans731; and SB 179 creates and funds the Nuclear Energy Development Grant Program, to be administered by the Kentucky Nuclear Energy Development Authority.732
• Maryland: A key provision of the Next Generation Energy Act (HB 1035 / SB 937) directs Maryland’s Public Service Commission (PSC) to create a procurement process for nuclear generation capacity.733 Modeled on the state’s procedures for procuring offshore wind generation, it is the first state policy created for financing nuclear reactor construction costs in a merchant power market. Upon receipt of a petition for a nuclear energy project, the PSC must undertake an application process, soliciting proposals for nuclear energy projects and selecting one or more that satisfy an undisclosed, yet-to-be-determined electricity price limit. Unlike CWIP, the law does not permit recovery of financing costs before the generation facility produces electricity, but it would provide a guaranteed rate for up to 30 years. The statute requires the PSC to conduct at least two more such application periods for new nuclear generation before 2031.
• North Dakota: HB 1025 authorizes up to US$300,000 from the state treasury for a study of “the feasibility, siting, and deployment of advanced nuclear power plants”734; and SB 2159 removes a prohibition on the state’s Energy Research Center at the University of North Dakota from researching or conducting projects resulting in the “exploration, storage, treatment, or disposal of high-level radioactive waste above ground.”735
• Tennessee: HB 1143 defines “clean or green energy or renewable energy” to require the inclusion of specified energy sources, including nuclear power—thereby barring municipalities from excluding nuclear power from their “clean or green energy” policies.736 SB 758 authorizes municipalities’ industrial development boards to acquire brownfield sites for nuclear power plant development.737
• Texas: HB 14 establishes the Texas Advanced Nuclear Energy Office and two grant programs that it would administer to support development of new reactors and nuclear infrastructure. Grants would range from up to US$12.5 million for project planning and development and up to US$120 for permitting, licensing, and construction. In total, the legislature authorized up to US$350 million for the program.738
• Utah: HB 249 establishes the Nuclear Energy Consortium and the Utah Energy Council to research and direct priorities for nuclear energy development, designates Energy Development Zones, and creates an Energy Development Investment Fund to provide tax increment financing incentives for energy projects.739 In addition, a concurrent resolution (HCR 5) calls on the U.S. Congress to enact legislation to reform permitting and environmental reviews to “expedit[e] energy infrastructure deployment,” including nuclear power.740

Industry Restructuring and Emerging Business Models

Since WNISR2023, we have detailed intersecting trends related to ownership changes and business strategies within the U.S. nuclear industry. Following publication of WNISR2024, there were several high-profile announcements of deals between major IT firms and nuclear power companies. However, incoming President Donald Trump immediately began efforts to roll back executive orders and policies enacted under the Biden administration. Despite Trump’s support for nuclear energy in his first administration and his inclusion of uranium-based fuels as a national priority in the energy emergency declaration issued on the day he took office, threats to repeal major energy programs in the Inflation Reduction Act created uncertainty about the fate of key financial incentives the nuclear industry is counting on to support major investment decisions in new reactor construction, license extensions, power uprates, and hydrogen co-generation.741

The industry became worried. Trade journal Nuclear Engineering International’s chief editor David Appleyard stated in May 2025 that for the nuclear sector Trumps budget proposals “represent sobering reading”:

There is no question that, even accounting for greater efficiency and better targeting of federal investment, losing hundreds of millions of dollars in annual funding will have dire consequences. (…) For nuclear the hope must be that the budget proposals form part of the on-going mission to ‘break things and move fast’ before eventual retrenchment and a more considered approach. The alternative is that even as much of the world looks to the US for some leadership and future thinking on nuclear, it is actually just breaking things.742

Furthermore, the administration’s promotion of fossil fuels, rejection of climate change, and repeals of climate policy measures have realigned the interests of nuclear generation companies and key industrial customers. For instance, the Biden administration began pressing the IT industry to reduce the emissions impacts of energy-intensive Artificial Intelligence (AI) data centers in 2024743; just months later, the industry is under no such pressure from the Trump administration, which even frowns upon emissions reductions as a business priority. As a result, utilities planning to serve gigawatt-scale loads from expected rapid growth in hyperscale data center operations are sidelining climate as an energy procurement consideration.744

Consequentially, the pace of new ventures slowed in the early months after Trump’s inauguration. With respect to subsidy and financing programs, the availability of federal production and investment tax credits and loan guarantees vitally affects nuclear industry business plans. Significantly, the one major corporate restructuring deal announced since mid-2024 involves large addition of fossil fuel generation: Constellation’s US$16.4 billion deal to acquire Calpine, the largest gas power generator in the U.S. If regulators approve the deal, Constellation will become the largest power generation company in the country, with 60 GW of generation capacity.745 Constellation CEO Joe Dominguez was quoted as saying that the Calpine acquisition is intended to capture market share in the data center power supply market:

The hyperscalers and other data providers that want to store AI data do not have enough carbon-free power options today… They’re going to have to use some natural gas, and we want to be able to participate in that.746

Those considerations notwithstanding, several developments in recent months could have significant implications for nuclear energy development in the U.S.:

• On 23 May 2025, President Trump issued four executive orders to promote an expansion of nuclear energy.747 They include setting a target to quadruple nuclear generation capacity by 2050; setting a 5-GW target for power uprates of existing reactors;748 directing the NRC to expedite licensing proceedings, with a limit of one year for license extensions and 18 months for reactor construction permits and licenses;749 expanding domestic uranium mining, enrichment, fuel production, and reprocessing capacity; directing the Department of Defense and Department of Energy to initiate construction of ten reactors on federal lands by 2030, outside of NRC licensing review, in order to power military bases and other national security infrastructure, including artificial intelligence data centers.750
• Congress enacted a bill with sweeping amendments to federal tax policies and budgets, which includes a rapid phaseout of tax credits for wind and solar generation.751 It leaves tax credits for nuclear energy largely unaffected, only adding restrictions on the eligibility of reactors owned by certain countries (including Russia and China) and the use of nuclear fuel produced by those countries.
• Following issuance of the executive orders, Westinghouse reportedly is meeting with the White House, proposing to provide the ten reactors that the Defense and Energy departments are directed to build, at a cost of US$75 billion.752
• Fermi America, a company recently co-founded by former Energy Secretary Rick Perry (during President Trump’s first term), announced a proposal to build four Westinghouse AP-1000 reactors on land at a Department of Energy nuclear weapons plant in Amarillo, Texas.753 On 7 July 2025, NRC issued a letter acknowledging Fermi’s “initial submittal of a Combined License Application (COLA).”754

The concrete implications of these developments are yet to be determined. Some measures covered in the executive orders will be dependent on fiscal appropriations by Congress, including expansions of nuclear fuel infrastructure and construction of reactors by the Defense and Energy departments.

AI, Data Centers, and Nuclear Power

There has also been a significant development affecting the trend of co-location and power supply deals between nuclear generators and data center developers. As reported in WNISR2024, Amazon Web Services (AWS) agreed to pay Talen Energy US$650 million for the rights to a large data center development planned at Talen’s two-unit 2.5-GW Susquehanna Nuclear Power Plant in Pennsylvania.755 The data-center campus includes options to expand in subsequent phases to consume as much as 960 MW, 40 percent of Susquehanna’s capacity.

A feature of the development included an agreement with grid operator PJM Interconnection that would have permitted AWS to classify an additional 180 MW of its data-center load (up to 480 MW) as located “behind-the-meter,” enabling Amazon to avoid costs associated with grid interconnections, transmission system upgrades, and approvals.756 Transmission owners Exelon and American Electric Power (AEP) filed a complaint with the Federal Energy Regulatory Commission (FERC) to void the deal and require AWS to go through the interconnection process, arguing that the data center will continue to draw power from the grid during times when Susquehanna is offline, and that other consumers will bear up to US$140 million in transmission and generation costs affected by AWS’s load.757 In November 2024, FERC ruled in the utilities’ favor, voiding the settlement.758

The decision had the effect not only of limiting the exemption for AWS to the first 300 MW of its data center load, as previously approved; but of signaling that FERC would view future data center colocations at existing power plants similarly. Talen appealed the decision, but FERC denied the appeal and affirmed the ruling in April 2025.759 In May 2025, Talen filed a further appeal in its suit against FERC in the U.S. Court of Appeals for the 5th Circuit.760 However, in June 2025, Amazon and Talen announced that they have amended their agreement and converted it into a conventional Power Purchase Agreement (PPA) for up to 1920 MW of Susquehanna’s generation capacity. The PPA is no longer specific to the co-located data center but is for the purchase of power for Amazon’s data centers in the PJM service area. The parties state that the agreement no longer rests upon the proposed exemptions from interconnection fees and, therefore, no longer requires FERC approval. Talen has not announced whether it will continue its legal action against FERC.

Following FERC’s denial of the appeal, Constellation announced that it had adjusted its data center strategy: it will pursue PPAs rather than colocation deals.761 Constellation had not reached any colocation agreements to date despite expressing interest in doing so already in 2022762; in fact, it is consistent with two prior announced deals with Microsoft. In 2023, Constellation stated it had reached an agreement with Microsoft for a “Virtual Hourly Matching PPA” (VHM PPA), to sell power to the tech giant’s Virginia data centers from its nuclear power plants in the region.763 A platform jointly developed by the corporations would ostensibly match transactions for power generated by Constellation reactors to Microsoft’s data center loads.764 Without further mentioning that VHM PPA, in September 2024, the corporations announced the deal for a 20-year PPA tied to the proposed restart of Three Mile Island-1 in around 2028, through which Microsoft would procure power for a number of its data centers connected to the PJM grid.765

Following the Three Mile Island restart-deal, major technology companies announced deals to procure nuclear energy. All of these other deals are with startup firms designing new reactors, for future procurement of power. Only two tech companies have made any direct investments as a result of any of the deals: in October 2024, Amazon led a group of investors investing a total of US$500 million in X-energy.766 At the same time, Amazon agreed to contribute a reported US$334 million to Energy Northwest (ENW) for a feasibility study of the utility’s proposal to develop up to three (320 MW) X-energy Xe-100 SMR plants; under that MoU, Amazon would agree to procure power from the first four-reactor Xe-100 plant.767 A third procurement MoU which Amazon announced with Dominion Energy to develop X-energy SMRs in Virginia is not reported to include any direct investment.768 In June 2025, AI microchip producer Nvidia invested an unspecified amount in Terrapower’s “$650 million fundraise.”769

Of other tech majors, Google announced MoUs with SMR developer Kairos Energy (500 MW) in October 2024770 and nuclear project developer Elementl (1.8 GW) in May 2025771; Meta issued a request for proposals for up to 4 GW of new nuclear capacity in December 2024,772 and announced a deal with Constellation to purchase zero-emissions credits from the 1120-MW Clinton-1 reactor in Illinois beginning in June 2027, after the state’s ZEC program ends.773 Amid those announcements, Oracle reportedly stated to investors in September 2024 that it has secured “building permits for three small modular nuclear reactors (SMRs),”774 though NRC has issued no such permits and the company has not provided any details to clarify or substantiate the claim.

In addition, Oklo announced deals with data center developers Equinix (500 MW) and Switch (12 GW).775 The Equinix deal included a “US$25.0 million prepayment made by Equinix to Oklo for the supply of power by Oklo.”776 In March 2025, Oklo issued a letter to its shareholders stating that it has “upgraded” its reactor design to 75 MW and providing additional details on the structures of its power supply deals.777

Since it submitted a license application to the NRC in 2020, Oklo has increased the capacity of its Powerhouse reactor from a 1.5-MW microreactor, to a 10-MW to 50-MW option it announced in 2022, to a 75-MW design in 2025. The NRC rejected Oklo’s 2020 license application in 2022, after the company failed to submit additional information required by the agency to complete its review.778 Oklo has been in pre-application discussions with NRC since 2022 and says that it intends, as of May 2025, to submit a new license application in late 2025.779 However, according to the information provided to investors in March, it is evident that Oklo’s power deals are not dependent on licensing or building reactors.780 The company reports that its deals are in the form of “master power agreements”, which apparently only obligate Oklo to deliver power to its customers, regardless of how or where it is generated. According to the shareholder letter, Oklo has contracted with a third company, RPower, a provider of gas-fueled generation, to generate the power for Oklo’s customers:

Oklo and RPower are partnering on a first-of-its-kind phased energy strategy to deliver immediate power and long-term clean energy for customers. RPower’s natural gas generators will provide bridge power for select projects, then shift to providing backup power as Aurora powerhouses come online, providing high uptime and reliability.781

Oklo did not report whether its power deals require it to complete the transition to supplying nuclear-generated electricity by any particular date, nor whether there are any penalties or contingencies if it does not ultimately succeed in doing so.

There have been limited developments with respect to hydrogen production and other cogeneration applications since mid-2024. Dow Chemical’s construction permit application for the Long Mott Generating Station clarifies that the four-reactor facility would have sufficient capacity to meet all of the steam supply needs of its adjacent UCC Seadrift chemical production plant, in addition to providing about 290 MW of net electricity generation.782 In early January 2025, the Internal Revenue Service issued tax rules implementing the Hydrogen Production Credit (Hydrogen PTC), which significantly expanded the capability for existing nuclear generation to qualify.783 The IRS rules still capped the eligible capacity from existing reactors for hydrogen production at 200 MW. Together with the uncertain fate of the Hydrogen PTC under the Trump administration, Constellation and other nuclear generation companies have not confirmed plans to proceed with hydrogen production.

Overview – Nuclear Power Contracts with Data Centers

Since 2021, nuclear power corporations and data center developers have announced plans to supply up to 35 GW of nuclear generation capacity to power artificial intelligence, cryptocurrency mining, and other hyperscale data centers.

Major Data Center Developers

Amazon

Four deals include power purchase agreements (PPAs) and Memoranda of Understanding (MoUs) for up to 7.2 GW of nuclear generation capacity:

◼︎ Talen (1920 MW): In March 2024, Amazon bought the rights to develop up to 960 MW of data center capacity at Talen Energy’s Susquehanna Nuclear Power Plan.784 In June 2025, the companies amended and expanded the PPA to 1920 MW of power for Amazon’s data centers in the PJM service territory.785

◼︎ X-energy (~5 GW): In October 2024, Amazon led US$500 million of investments in X-energy, which included an MoU to develop “more than 5 gigawatts [GW]” of Xe-100 SMRs by 2039.

◼︎ Energy Northwest (320 MW): In October 2024, Amazon entered an MoU with Energy Northwest, which includes US$334 million contribution by Amazon toward a feasibility study. The agreement gives Amazon the rights to power from the first of three X-energy SMR power stations (320 MW each) that Energy Northwest proposes to build.

◼︎ Dominion (300 MW): In October 2024, Amazon entered an MoU with Dominion Energy to procure 300 MW of power from an unspecified SMR to be built at the North Anna Nuclear Power Plant.

Alphabet (Google)

Two deals include MoUs for up to 2.3 GW of nuclear generation.

◼︎ Kairos (500 MW): In October 2024, Alphabet announced an MoU with molten salt reactor developer Kairos to build up to 500 MW of 70-MW SMRs by 2035.786

◼︎ Elementl (1800 MW): In May 2025, Alphabet and project developer Elementl announced an MoU for three nuclear projects, totaling 1800 MW.787 The announcement did not specify reactor designs or locations.

Meta (Facebook)

One contract and one solicitation, totaling up to 5.1 GW of nuclear generation.

◼︎ 1-4 GW solicitation: In December 2024, Meta issued a request for proposals (RFP) for up to 4 GW of nuclear generation capacity.788 Meta reports it is evaluating “over 50 qualified submissions.”789

◼︎ Constellation (1120 MW): In May 2025, Meta and Constellation announced a contract for the sale of zero-emissions credits (ZECs) from the Clinton Nuclear Power Plant in Illinois.790 The contract begins in June 2027, after Illinois’ state ZEC program expires.

Microsoft

Two MoUs with Constellation, for at least 835 MW of nuclear generation.

◼︎ Three Mile Island-1 restart (835 MW): In September 2024, Microsoft and Constellation announced an MoU for a 20-year PPA for the sale of power from the TMI-1 reactor, which was retired in 2019.791 The term of the PPA would begin when Constellation brings the reactor out of retirement, projected for 2028.

◼︎ Virtual Hourly Matching PPA: In June 2023, the companies announced a PPA to purchase power from Constellation reactors in the PJM service territory for Microsoft data centers in Virginia.792 Neither the announcement nor Constellation’s filings with the Securities Exchange Commission included relevant details on the PPA, neither the amount of power, nor the duration of the contract or the price.793

Oracle

In September 2024, data center developer Oracle reported to shareholders that it has “building permits” to build three SMRs, totaling 1 GW of capacity.794 Founder, chairman, and Chief Technology Officer Larry Ellison provided no further details. The NRC has not issued construction permits for any SMR projects, so the factual basis for the corporation’s statements is unclear.

Nvidia

◼︎ In June 2025, Microchip manufacturer Nvidia invested in a US$650 million funding round in TerraPower, developer of the 345 MW Natrium sodium-cooled fast reactor proposed in Wyoming.795

Minor Data Center Developers

Smaller players in the data center industry have also signed PPAs and, in some cases, co-location agreements with nuclear power plant owners.796

Mawson Infrastructure Group (153 MW)

◼︎ Mawson currently operates 153 MW of modular data centers on retired steel plant sites near the Beaver Valley and Perry nuclear power plants.797

NE Edge, LLC (300 MW)

◼︎ In 2023, developer NE Edge secured the rights to develop a 300-MW798 data center at Dominion Energy’s Millstone-2 and -3 reactor site in Connecticut.799 Construction of the data center has not begun, pending Dominion’s submission and approval of an application to the Connecticut Siting Commission.

Standard Power (112 MW)

◼︎ In 2021, Standard Power entered into PPAs with Energy Harbor (now owned by Luminant) to power its data centers in Ohio.800 As of 2024, the company was operating 112 MW of data centers at sites in Coshocton and Conesville, OH.801 The company reportedly plans to expand the Conesville site with another 1165 MW of data centers under further PPAs with Luminant’s nuclear power plants in Ohio and Pennsylvania.

Nuclear Reactor Developers

Three other corporations have announced power contracts or development projects to power data centers with nuclear reactors.

Fermi America (4.4 GW)

In June 2025, the new company Fermi America, led by former Texas Governor and U.S. Energy Secretary Rick Perry announced plans to develop the Donald J. Trump Advanced Energy and Intelligence Campus, with up to 11 GW of generation capacity for power data centers, all to be located at the Pantex nuclear warhead facility in Amarillo, TX. Fermi America filed an “initial combined operating license submission” to build four Westinghouse AP-1000 reactors on 17 June 2025.802

Last Energy (600 MW)

Microreactor design company Last Energy has announced plans to construct thirty 20-MW reactors (totaling 600 MW) at a site in Haskell, Texas, to sell power to data centers in Texas.803 It initiated pre-application engagement for an early site permit (ESP) with the NRC in February 2025.804

Oklo (14 GW)

Oklo has announced “master power agreements” with four customers, totaling 14 GW of generation capacity. With the exception of oil and gas producer Diamondback Energy, the deals are with data center developers:

◼︎ Switch (12 GW)805

◼︎ Equinix (500 MW)806

◼︎ Prometheus Hyperscale807 (100 MW)808

According to a letter it provided to its shareholders, Oklo will initially serve its master power agreements with gas generators provided by a third party, Rpower.809

Criminal Investigations of Nuclear Power Corporations

Since 2017, the U.S. Justice Department has opened three separate investigations against utility corporations over criminal activities related to nuclear power. The cases have resulted in indictments of corporate executives, lobbyists, and state officials. The cases have been accompanied by additional lawsuits and state-level investigatory proceedings, and they have had political ramifications which appear to have had further impacts on the industry, economically, as well as legally and politically. For detailed background on these cases see WNISR2023 and WNISR2024. There have been significant developments in all three, in recent months:

SCANA-Westinghouse fraud case. The last remaining criminal indictment was resolved in December 2023, when Jeffrey Benjamin, former senior vice president for new plants and major projects at the Westinghouse Electric Company—who directly supervised all new nuclear projects worldwide during the V.C. Summer project—pleaded guilty for his role in covering up mounting delays and cost overruns during the construction of the V.C. Summer-2 and -3 reactors, which were canceled in 2017 after over US$9 billion were spent and nine rate increases for electricity customers.810 On 20 November 2024, Benjamin was sentenced to one year and one day in prison and a fine of US$100,000.811

Exelon-Commonwealth Edison corruption case. The final suspect to be indicted in the corruption case involving nuclear bailout and utility legislation in Illinois is former Speaker of the state House of Representatives Michael Madigan. The trial was originally scheduled to begin in April 2024, but was postponed until October 2024 in deference to a pending U.S. Supreme Court decision in a criminal case involving so-called “gratuities” paid to elected officials, which may have superseded the prosecution’s case against Madigan.812 On 12 February 2025, a jury convicted Madigan on ten of the 23 charges against him, with acquittals on seven charges and six ending in mistrial.813 Under each of the three counts of wire fraud charges of which he was convicted, Madigan faced up to 20 years in prison; prosecutors initially also sought US$3.1 million in restitution for the bribes he accepted, but later dropped that request.814 On 13 June 2025, Madigan was sentenced to 7.5 years in prison and US$2.5 million in fines; the judge apparently enhanced the sentence finding that Madigan had “perjured himself repeatedly in his trial testimony.”815

FirstEnergy corruption case. Former Speaker of the Ohio House of Representatives, Larry Householder, and three lobbyists are serving prison sentences for their role in a bribery and corruption scandal that unfolded between 2017 and 2019. FirstEnergy gave Householder, his associates, and his political action committee, Generation Now, close to US$60 million to pass HB 6, a bill that provided a US$1.05 billion “taxpayer-funded bailout” to two reactors that FirstEnergy owned at the time. Householder and co-conspirator Matt Borges, former Ohio Republican Party Chairman, appealed their conviction, but the court denied the petition in May 2025.816

The case has continued to unfold, with further federal and state indictments since December 2023. That month, federal prosecutors brought charges against Sam Randazzo, the former Chairman of the Public Utility Commission of Ohio (PUCO), who allegedly accepted US$4.3 million from FirstEnergy and helped to draft and implement HB 6.817 In February 2024, Ohio Attorney General Dave Yost filed charges against Randazzo and two former FirstEnergy executives, then-CEO Chuck Jones and then-Vice-President Michael Dowling.818 In April 2024, Randazzo died by suicide.819 He is the second suspect indicted in the case who has taken his own life before going to trial, following the death of lobbyist Neil Clark in 2021 subsequent to his indictment.

Yost’s prosecution of Jones and Dowling is scheduled to go to trial on 26 January 2026.820 In addition, on 17 January 2025, a federal grand jury indicted the former executives for “participating in a racketeering conspiracy.”821 Dowling reportedly plans to call Governor Mike DeWine and Lieutenant Governor Jon Husted as witnesses.822 In June 2024, it was reported that text messages between DeWine and Jones suggest the governor may also have been involved in the bribery scheme. The text messages allegedly show that DeWine asked Jones for a significant campaign donation a month before the 2018-election, and messages between Dowling and Jones in 2019 would state that DeWine actively worked to get the state legislature to pass HB 6.823

In April, a court ruling compelled testimony from three former FirstEnergy lobbyists and a former executive, enabling an investigation by the Public Utilities Commission of Ohio to move forward.824

Conclusion

The number of reactors and annual nuclear generation again increased slightly in 2024, with Vogtle-4’s first grid connection resulting in a minimal increase of less than 1 percent in total annual nuclear generation. Despite the startup of both new Vogtle units in 2023–2024, the average age of the U.S. reactor fleet continues to advance, reaching 43.7 years as of mid-2025. Efforts to restart retired reactors expanded somewhat, with two further proposals joining that of Palisades-1: Duane Arnold-1 and Three Mile Island-1, which were retired in 2020 and 2019, respectively. Palisades-1 remains the most advanced proposal, though its timeline, cost, and ultimate prospects have been clouded by inspection results finding that nearly 1,000 steam-generator tubes are cracked and must be plugged. In addition, the proposal to continue operation of Diablo Canyon-1 and -2 has moved forward, with review of the license renewal application underway at the NRC. Further, the industry has announced plans to seek subsequent license extensions (from 60 years of operation out to 80 years) for more than half of currently operating reactors, and to seek power uprates of at least 16 reactors in the coming years.

Vogtle-4’s first grid connection in March 2024 means there are no commercial power reactors under construction in the U.S. for the first time since 2007 (official date of Watts Bar-2 construction restart). No utility companies have made final investment decisions to build new reactors of any type, and the cancellation of NuScale’s Carbon Free Power Project in November 2023 has shaken confidence that new designs for SMRs and non-LWRs (Light Water Reactors) will break the pattern of cost escalation and construction delays. Three construction permit applications are currently under review: TerraPower’s Natrium project in Wyoming, a sodium-cooled fast neutron design; Dow Chemical’s Long Mott Generating Station, comprising four X-energy Xe-100 graphite-moderated, high-temperature, gas-cooled SMRs; and TVA’s Clinch River SMR project to build a GE-Hitachi BWRX-300 reactor.

Executive orders issued by President Donald Trump and legislation enacted by Congress in May and July 2025, respectively, could have significant impacts on new reactor construction in the coming years. The legislation largely preserves federal subsidies for nuclear power plants enacted in 2022, while rapidly phasing out subsidies for wind and solar generation. Among other measures, the executive orders call for a quadrupling of nuclear generation capacity by 2050, direct the NRC to rapidly accelerate review of license applications, and direct the departments of Defense and Energy to initiate construction of ten new large reactors by 2030. In response, the NRC has issued statements committing itself to comply with Trump’s licensing directives, and companies have announced proposals for new reactor projects: Westinghouse has reportedly met with government officials, offering to construct ten AP-1000 reactors at a projected cost of US$75 billion; and a new company, Fermi America, has submitted an “initial license application” to build four AP-1000 reactors to power a large data center complex at a Department of Energy plutonium production plant.

Corruption and fraud investigations involving both new reactors and subsidies for operating reactors continued, with the final prosecution in one of them concluding with the sentencing to federal prison of former Westinghouse Vice President Jeffrey Benjamin in the V.C. Summer fraud case. Former Illinois House Speaker Michael Madigan was convicted on ten corruption charges for accepting bribes from Exelon; he was sentenced in June 2025 to 7.5 years in prison and US$2.5 million in fines. In the US$61 million FirstEnergy bribery case, federal and state indictments of former FirstEnergy executives, ex-CEO Chuck Jones and ex-VP Michael Dowling are moving forward, with the state trial scheduled for January 2026.

Fukushima Status Report

Overview of Onsite and Offsite Challenges

Fourteen years after the disaster began at the Fukushima Daiichi nuclear power plant, Tokyo Electric Power Company (TEPCO) for the first time removed bits of radioactive debris from a heavily contaminated unit—a significant milestone in the plant’s decommissioning that at the same time underscores how immense the task ahead is. The two removals took place three years later than initially scheduled and gleaned a total of just 0.9 grams of material, out of an estimated 880 tons that need to be cleared out. There is no plan yet for how to remove the bulk of the material.

Offsite, decontamination efforts are inching along, with the government aiming to complete around summer 2025 a plan to store or reuse the millions of cubic meters of radioactive soil removed from contaminated areas thus far. Japan appears to have largely kept its food supply free of radioactive contamination and most countries have now lifted bans on Japanese agricultural and seafood imports. China, one of the holdouts, agreed last year to gradually resume seafood purchases. It partially lifted its ban at the end of June 2025. But a WNISR dive into Japan’s Food Contamination Monitoring reveals an opaque and apparently disorganized system that makes it difficult for outside observers to understand or feel comfortable with the data.

Onsite Challenges

Removing Highly Radioactive Nuclear Fuel Debris825

To understand just how massive the challenge of decontaminating and dismantling Fukushima Daiichi is, it is instructive to look at TEPCO’s struggle to remove its first pinch of fuel debris last year (see Table 12).

Fuel debris is the mess left after fuel assemblies in three reactor cores overheated and melted, boring through the steel pressure vessel that surrounded them, carving a path through layers of metal, wiring and electronic equipment and eating into the steel-and-concrete foundation in the primary containment vessel below.

The resulting debris, a hardened mix of nuclear fuel, metals, and minerals, is highly radioactive and thus continues to generate heat. It has to be kept cool—the regulatory limit for the bottom of the pressure vessel is 80°C826—to ward off any possibility that the nuclear fuel in the mixture could overheat again or start a thermal runaway reaction. TEPCO says that even without cooling, the debris temperature should plateau at around 240ºC, well below the 1,200˚C that would spur a thermal runaway reaction.827 TEPCO keeps the debris cool by maintaining a continuous injection of water into the reactors. For the past year, that has kept the temperature at the bottom of the pressure vessels of all three of the damaged units below 40°C.828 TEPCO also injects nitrogen into the damaged pressure vessels to prevent the buildup and combustion of hydrogen, something that had led to explosions in the days after the meltdowns.

TEPCO estimates there is around 880 tons of fuel debris in the three units. It is the source of the continuing radioactive contamination that seeps out of the plant through water, air, and dust. To end that threat and stabilize the onsite situation, the debris must ultimately be removed, packaged, and sealed off somehow from the outside world.

Debris Sampling

Step one in the debris removal process is taking a small sample and assessing its composition, in order to help gauge the radiation risks to workers and figure out how to remove larger quantities and store it safely at scale.829 TEPCO decided to start with Unit 2, since radiation levels inside its reactor building are relatively low compared to the other two units that experienced meltdowns—Units 1 and 3 —and the building itself is in relatively good shape, meaning there is less chance of scattering radioactive particles to the outside during removal.830

The primary containment and pressure vessels of all three units are far too radioactive for human workers to enter for any length of time, with lethal dose-rate levels estimated at between several Sieverts per hour (Sv/h) to several hundreds of Sv/h, thousands of times the regulatory annual dose limit.831,832 But TEPCO has been surveying the inside of the vessels with robots and had more information on the location of the debris in Unit 2 than for the other units.

Initially, the sampling was supposed to have started by 2021 but was delayed until the fall of 2024 by the COVID-19 pandemic and unexpected problems in gaining access to the containment vessel.833 TEPCO and partner engineering companies decided the best way to take the first sample was with a piece of equipment shaped like a fishing pole. To get to the debris, engineers guided a telescoping rod into a narrow shaft in Unit 2’s containment vessel that was normally used for equipment installation. The rod then extended and lowered a cable equipped with a camera, light, and grasping pincers to the bottom of the vessel to take the sample.834 The whole process was expected to take three weeks as the equipment navigated obstacles.835

There were problems. Workers installing the sampling equipment in the Unit 2 reactor building had to wear full protective gear and were restricted to 15- to 30-minute shifts in order to limit radiation exposure. In August 2024, they realized they had installed sections of the rod in the wrong order, leading to a redo. They also had to replace two cameras,836 which failed in the highly radioactive environment.837

Eventually, TEPCO extracted 0.7 grams of debris on 7 November 2024,838 and another 0.2 grams from a different spot closer to the center of the vessel on 23 April 2025.839 There is around a billion times that amount left to take out from the three reactors.

What Happens Next

Now, as research labs analyze the composition of the sampled debris—a process expected to take 12 to 18 months840—TEPCO is also mulling plans for removing the bulk of the material, much of it far more radioactive than the samples. The company is planning to keep pecking at the debris pile in Unit 2, testing and honing the equipment and trying to figure out ways to take out more than just a few grams at a time.841 For its next tries, TEPCO is hoping to use a robotic arm for the sampling, rather than pincers hanging from a cable, since it is able to move around more and can take out greater quantities. TEPCO is aiming for a first trial of the arm in the second half of FY2025.842

The company is also slowly building structures that will let it remove the 615 intact spent-fuel assemblies sitting in a cooling pool atop the Unit 2 reactor building. TEPCO aims to start removing the Unit 2 spent fuel by FY2026 and the 392 spent-fuel assemblies from the highly contaminated Unit 1 by FY2027–28.843 TEPCO moves fuel assemblies from the pools atop each unit to a common pool. When they have cooled enough, the assemblies are transferred to dry storage casks in a temporary facility onsite.844

TEPCO finished removing the 1,456 spent-fuel assemblies from the relatively unscathed Unit 6’s spent-fuel pool on 16 April 2025, leaving 428 new-fuel assemblies at the unit, 198 in the spent-fuel pool and 230 in a storage facility in the reactor building.845 The company plans to ship 56 of those assemblies back to their manufacturer in the U.S. starting in the latter part of FY2025.846

TEPCO plans to start removing fuel assemblies from Unit 5 in July 2025. Units 5 and 6 are located about 500 meters from Units 1–4. TEPCO completed unloading the spent-fuel assemblies from Unit 3 in 2021 and Unit 4 in 2014.847 Units 4, 5, and 6 were not operating when the tsunami hit and did not suffer fuel meltdowns. TEPCO aims to have all the fuel assemblies removed from Units 1–6 by 2031.

TEPCO’s success in removing the spent-fuel assemblies from Unit 3 is making that reactor the prime candidate for the first large-scale debris removal. TEPCO is looking at three potential methods (see Japan Focus in WNISR2024) and is expected to finish the studies by the middle of FY2025.848 The biggest challenge is figuring out how to shield workers and the outside world from extreme levels of radiation while shoveling tons of debris past obstacles like broken equipment and damaged metal scaffolding, then into storage. The latest mid-to-long-term decommissioning plan envisions that work starting some time before 2036.849

The vision is to finish debris removal and the rest of the decommissioning work by 2051, a goal Prime Minister Shigeru Ishiba reaffirmed on the 14th anniversary of the Fukushima events in March 2025.850 Given the difficulties ahead and TEPCO’s track record so far, many observers are skeptical.

“With this debris sample removal, we move into the third [and final] stage of the [decommissioning] process,” the Nihon Keizai Shimbun wrote in a 13 November 2024 editorial, “But no expert thinks we will be able to revert the land back to its untouched state in 27 years.”851

Contaminated Water852

TEPCO is slowly making headway in the battle against contaminated water at the Fukushima Daiichi compound, although gains remain hard-fought and incremental. In the latest sign of progress, TEPCO in FY2024 released to the sea twice as much treated water as the site generated in newly contaminated water, according to WNISR calculations. In February 2025, the company started dismantling some of the 1,000-plus tanks—most containing 1,000 tons or more of contaminated water—in order to make room for facilities to eventually store fuel debris, although it only has firm plans to remove 12 tanks this fiscal year.853

Contamination largely results when water injected to cool the reactors comes into contact with the highly radioactive fuel debris and leaks into the basements of the buildings, where it mixes with groundwater or rainwater and seeps out into the environment.

To keep that contamination from spreading, TEPCO has installed a seawall and a massive underground wall of ice that surrounds the damaged units. It took steps to keep groundwater around the units relatively low and clean by digging wells to pump out water and covering around 1.4 million m2 of land with asphalt.

Those steps have helped TEPCO bring down the amount of contaminated water generated each day from 540 m3 in May 2014 to an average 70 m3 in FY2024, and 50 m3 in the first quarter of 2025.

Still, the amount of contaminated water on the site continued to increase until August 2023, when TEPCO started its controversial step of releasing into the ocean water filtered to remove radioactive elements aside from tritium, a radioactive isotope of hydrogen. In FY2024, TEPCO released 55,000 m3 of treated water,854 while the site generated around 25,550 m3 of contaminated water, calculating from the daily average reported by TEPCO.

By 28 April 2025, TEPCO had discharged a cumulative 94,000 m3 of treated water. That left roughly 1.3 million m3 of contaminated and treated water on site as of 19 June 2025,855 stored in nearly 1,100 tanks.856 TEPCO plans to treat and discharge all the contaminated water that collects on the site by the targeted decommissioning deadline of 2051.

A similar volume of water as in the previous year but with a 20 percent higher tritium content

For FY2025, TEPCO plans to release around 54,600 m3 of water, representing roughly 15.3 trillion becquerels (Bq)857 of tritium—a similar volume of water as in the previous year but with a 20 percent higher tritium content.858

The water-treatment system, Advanced Liquid Processing Systems (ALPS), continues to underperform, with only 34 percent of the 1.18 million m³ of the water it had processed satisfying regulatory standards as of March 2025—little changed from the previous year. The rest needs to be repurified (see Figure 52).859

1. Percentages of Treated Water and Water to be Re-purified

Source: TEPCO Contaminated Water Portal Site, May 2025

Worker Safety

Worker radiation exposure remains reportedly low overall, with an average exposure of 2.08 mSv/person for FY2024—far below the regulatory limit of 100 mSv over a five-year period and 50 mSv annually, as well as TEPCO’s own annual limit of 20 mSv.860

However, in a worker survey conducted between September and October 2024, concerns about contamination—particularly bodily contamination—increased compared to the previous survey.861 TEPCO speculates this might have to do with an accident in October 2023, when workers cleaning ALPS pipes were sprayed with radioactive water, resulting in contamination levels high enough to send two people to the hospital. The Ministry of Economy, Trade and Industry said after their discharge that “there were no physical problems with any of the workers, and no trouble were observed at the skin of the contaminated area.”862

Other worker-safety concerns are accidents and heat stroke. Fifteen people were injured in workplace accidents at Fukushima Daiichi in FY2024, one seriously enough to require a leave of absence of more than two weeks. Eight people suffered heat stroke or other heat-related stress. Both those numbers are up from the previous fiscal year.863 On 13 June 2024, one worker was found collapsed in a rest area in the compound; he was taken to a hospital and confirmed dead. TEPCO said there was no radioactive contamination on his body and the cause of death was unclear.864

Offsite Challenges

Approximately 165,000 people evacuated from Fukushima Prefecture after the nuclear disaster began. As of 1 February 2025, the prefecture still officially recorded 24,644 evacuees, 19,673 of them living outside Fukushima.865 The total number of people who have returned to former evacuation zones is uncertain and varies widely between locations.

Decontamination work has allowed the government to chip away at Fukushima Prefecture’s “Difficult-to-Return Zones”—areas considered hazardous to live in because annual radiation dosage levels are higher than 50 mSv per year, and are likely to remain above 20 mSv per year even after five years. On 31 March 2025, the government lifted the designation on pieces of Iitate and Katsurao villages,866 leaving around 2.2 percent of the prefecture still categorized as “Difficult-to-Return Zones”, as of December 2024.867

Disposal of Contaminated Soil868

The soil, leaves, timber, and other waste removed in decontamination efforts—mostly in Fukushima Prefecture—is currently stored in interim facilities in two towns bordering the Fukushima Daiichi power plant site. There was around 14 million cubic meters of it as of the end of December 2024, enough to fill 5,600 Olympic swimming pools, even though much of the heavily forested prefecture remains untouched. See WNISR2024 for a discussion of the effectiveness of decontamination in Fukushima Prefecture.

The government is legally mandated to move that contaminated soil and waste out of Fukushima Prefecture by 2045. Japan’s Nuclear Emergency Response Headquarters, a group under the Cabinet Secretariat, created a new committee in December 2024869 to take responsibility for planning and implementing that removal, supported by the Ministry of the Environment. The committee released its basic policy for the removal and reuse of the contaminated soil in May 2025 and plans to put together a roadmap for implementation around summertime.870 Broadly, the government plans to use, e.g., in a range of construction applications, the soil that is judged safe for reuse, with radiation levels below 8,000 Bq per kilogram. Around three-quarters of the contaminated soil and waste falls into that category.

The Japanese government has decided the rest is to be shipped and stored permanently outside Fukushima Prefecture, in view of the already extreme burden of the nuclear accident on the area’s residents.871 In early 2025, the Ministry of the Environment released various scenarios for dealing with that heavily contaminated material, including processes for concentrating it into a smaller, more radioactive mass.

Between 2023 and 2024, the International Atomic Energy Agency (IAEA) reviewed the ministry’s approach for recycling and disposal of contaminated soil and deemed it “consistent with the IAEA Safety Standards.”872 The ministry has been conducting its own trials of reusing contaminated soil, including one in Fukushima where the soil was used as filler under rice paddies. The resulting rice had radiation levels well below Japan’s food limits, the ministry says.

However, the soil-recycling idea has gotten a lot of pushback from communities where other trials have been proposed, with activists raising concerns that dust from construction or other reuse projects could be inhaled by those nearby and lead to dangerous internal radiation exposure.873 In early 2025, Tokyo residents protested a plan to use contaminated soil under a flower bed in the popular Shinjuku Gyoen National Garden, while residents of Tokorozawa, Saitama denounced a trial planned for their city.874 The government has scrapped those trials, according to local media.875

At the end of May 2025, the government committee overseeing the contaminated waste removal proposed leading the way by using some of it at the prime minister’s residence, potentially under a garden; construction is reported to be starting in July.876

Meanwhile, an investigative piece in the Bulletin of the Atomic Scientists from January 2025 serves as a reminder of the challenges of tracking radioactive contamination and its effects. The article describes the discovery around 2016 that highly radioactive cesium microparticles had wafted to Tokyo following the Fukushima Daiichi meltdowns—and how the research detailing those findings was kept out of publication for three years amid academic infighting and wariness over a politically sensitive topic.877

Food Contamination Monitoring

One of the thorniest issues facing Japan following the Fukushima Daiichi nuclear accident has been controlling, monitoring, and communicating about potential radioactive contamination in its food supply.

Testing for radionuclides has shown vanishingly few food samples that exceed contamination limits—at least among cultivated plants and animals where farmers can better control the environment.878 When products do exceed those limits, the Japanese government pulls them off the market.

Most of the 55 countries and regions that initially halted imports of food from Japan have lifted those bans, with only China, Russia, South Korea, Taiwan, Hong Kong, and Macau still restricting some products as of mid-2025.879

China, which had banned all imports of Japanese seafood following Japan’s decision to start releasing treated water from the Fukushima nuclear plant into the ocean, agreed in September 2024 to gradually lift that ban in return for expanded cooperation in monitoring the releases.880 At the end of June 2025, China agreed to lift its import ban on seafood products from all but Fukushima and nine other prefectures, as long as exporters provided government certificates of origin and tests for a range of radionuclides.881

TEPCO has been discharging the water in batches since August 2023,882 and the IAEA has declared it “in line with international safety standards.”883 The European Union, Norway, Iceland, Switzerland, and Liechtenstein lifted their remaining food import bans that same month, shortly before the water discharges began.884 (See also section on Contaminated Water.)

Still, parts of the Japanese countryside remain significantly contaminated, especially forested areas and streams in Fukushima which are nearly impossible to scrub of radioactive elements. Wild game, fish, and plants from those areas are still showing high levels of radioactivity.

In FY2024, 29 percent of the wild boar meat from Fukushima Prefecture that was tested exceeded radiation limits,885 with one specimen caught in February 2025 in the heavily contaminated village of Futaba registering a radioactive cesium level of 13,000 Bq/kg, 130 times the legal limit of 100 Bq/kg.886 Scientists are still researching the ways in which different organisms, from mushrooms to fish, absorb radioactive elements.

Other prefectures such as Miyagi and Gunma also see high levels of contamination in boars and other wild game, while foraged plants and mushrooms from prefectures as far away from Fukushima as Nagano and Shizuoka have logged radioactivity levels exceeding limits during the past year.887

Japan’s radionuclide testing regime remains opaque and appears disorganized, making it challenging for the government to convince many international observers, and even its own citizenry, that it has control over the situation.

Contamination and Response

The meltdowns and explosions at Fukushima Daiichi spread radioactive particles over a wide and uneven swathe of the surrounding countryside, much carried northwest by winds. Japan has focused its attention on radioactive cesium, which was released in large quantities. In particular, it has concentrated on the isotope Cesium-137 which has a half-life of around 30 years and whose effects can linger for decades or even centuries.

Testing after the meltdowns found high levels of radioactive elements in products like spinach and milk from regions near the nuclear plant. It also found significant contamination in tea leaves hundreds of miles from the accident site and in cattle around the country that had eaten tainted rice straw, showing that radioactivity could be transferred in unexpected ways and pop up in surprising parts of the food chain.888 The Japanese government scrambled to establish contamination limits and radionuclide-testing programs to keep unsafe food off shelves and plates.

Japan’s current contamination limits went into effect in 2012, the year after the initial events. Like many other international guidelines set by groups such as the European Union or the Codex Alimentarius, a Joint Venture of the United Nations’ Food and Agriculture Organization (FAO) and the World Health Organization (WHO) that oversees food safety, Japan’s standards aim to limit the maximum exposure to radionuclides through food to 1 mSv per year for adults.889 That is a level considered appropriate by expert groups such as the International Commission on Radiological Protection,890 although scientists are still studying how low levels of radiation exposure affect the body, particularly over time.

For most foods, Japan caps the amount of allowable radioactivity at 100 Bq/kg, under an assumption that as much as 50 percent of what is eaten could be contaminated due to the Fukushima nuclear accidents. Drinking water, milk, and infant food have lower limits.

The E.U. caps levels of cesium in most foods at 1,250 Bq/kg and the UN-backed Codex Alimentarius at 1,000 Bq/kg, for a scenario involving the import of contaminated food comprising around 10 percent of total consumption. Both bodies note that guidelines may be revised if those assumptions change.

Japan’s Nuclear Emergency Response Headquarters, a group under the cabinet office, issues and updates guidelines for testing food products. The tests themselves are conducted by a variety of local governments, largely at the stage of production, before anything is sold or distributed. A Fukushima Prefecture-backed association of farmers, retailers, and consumers, for example, voluntarily tested all bags of rice grown prior to 2020—screening more than 11 million bags per year at one point.891 The blanket testing found 71 bags of rice that exceeded Japan’s radioactivity limit in 2012, 28 in 2013, two in 2014, and none in subsequent years.892 The prefecture still bans shipments or requires blanket testing of rice grown in areas that were more heavily contaminated.893

Items found to have exceeded the limits are pulled from distribution or consumption until subsequent tests show them to be clear.894

The government also conducts twice yearly tests of baskets of foodstuffs—chosen to mimic local diets—sold at retail outlets in 15 locations.895

Japanese farmers in affected areas have been attempting to lower radiation levels on their land by, for instance, spreading fields with fertilizers that limit cesium absorption by plants and using high-pressure sprays to cleanse trees in orchards.896

There have been only a handful of reported instances of contamination exceeding maximums for Japanese vegetables, fruit, livestock, and fish for the past decade, with most of those coming from Fukushima Prefecture.897

Puzzling Out the Data on Contamination

A closer look at Japan’s testing regime, however, reveals a system that appears messy and very hard to understand.

The government’s testing guidelines apply to the 17 eastern prefectures or municipalities in the vicinity of the nuclear disaster. Other areas and organizations conduct tests on a voluntary basis; some are included in the government’s overall testing tally and some are not.

The guidelines focus testing on food categories and locations where products have exceeded contamination limits and reduce monitoring for products that have shown little or no radioactive exposure. In the most recently revised guidelines, for instance, the items subject to testing are:

• products that have recorded cesium contamination levels of more than half the maximum during the past fiscal year;
• milk, beef, and log-grown mushrooms, which could be contaminated through feed or the wood on which they are cultivated;
• fish;
• some processed foods like dried mushrooms or seaweed; and
• products where restrictions were lifted in the past fiscal year or which will be produced for the first time following the accident, with no record of testing.898

In line with the focus on contaminated items, the government in 2017 started separating most products into those whose feeding and cultivation was “possible to manage”—generally farmed vegetables and livestock—versus “difficult to manage”—largely wild game, fish, and foraged plants, where the vast majority of high radiation readings were found.899

In FY2024, of the 193 items that exceeded contamination limits, 189, or 98 percent, were in the “difficult to manage” category, according to figures released by the Ministry of Health, Labour and Welfare (MHLW).900

The approach has led to an overall decline in the number of testing samples as decontamination of farmland progressed, from a peak of more than 340,000 in FY2015901 to just under 46,000 in FY2024902—with a sharp drop in FY2020, when four prefectures switched from blanket testing of cattle to random sampling, according to MHLW figures. The MHLW does not include results from Fukushima Prefecture’s blanket testing of rice.

The focus on contaminated items has also meant that the “difficult to manage” category is a growing proportion of the overall sampling data as testing concentrates there. In FY2020, around 60 percent of test samples were in the “possible to manage” category versus 34 percent that were in the “difficult to manage” category. By FY2023 that had flipped to 37 percent in “possible to manage” versus 58 percent in “difficult to manage.”903

In FY2024, the number of wild plants, mushrooms, and game samples shot up to more than 16,000, around 35 percent of the total.904

Still, the growing concentration of testing in the most contaminated categories makes overall contamination ratios difficult to parse: the share of food samples that exceeded contamination limits has actually been increasing as a percentage of the overall, by a factor of six from 0.06 percent in FY2019 to 0.42 percent in FY2024, according to MHLW figures.905

A Hodgepodge of Results Difficult to Interpret906

An even bigger challenge is that the testing guidelines are relatively vague, leaving details such as where and how much of various products are tested to the local governments, which implement them in different ways. That leads to a hodgepodge of results that are impossible to interpret without checking with each locality.

The local testing entities too are split up in odd ways and often do not integrate their reports. Fukushima Prefecture’s main radionuclide monitoring report for agricultural products, for instance, does not include results from two of the prefecture’s biggest cities, Koriyama and Iwaki, which do their own testing. Nor does it include tests of wild fish and game or processed foods, which are overseen by different prefectural divisions and reported separately.907 Until FY2020, it also did not include the results of most of the prefecture’s rice tests, which were reported separately before Fukushima largely switched to spot from blanket testing.908

The only comprehensive compilation of testing data is put out by the MHLW, which releases monthly and annual numbers broken down by prefecture and category of product. But it often takes months for the ministry to process and release test results, meaning the figures the ministry puts out may not match those released by the local governments, and may not reflect what was actually collected during each period.

Testing data released by the MHLW for FY2022, for instance, showed only 5,963 samples taken in Fukushima Prefecture, less than half the amount sampled in the previous or following years, an anomaly noted by WNISR. Fukushima Prefecture’s FY2022 radionuclide monitoring report for agricultural products, however, shows 11,208 samples, even though it reflects a smaller subset of tests.

An official overseeing the Fukushima report told WNISR that the MHLW’s delayed release of test results could explain some of that difference.909 For instance, one test result for beef that the prefecture had published on 1 August 2022 was not released by MHLW until 11 April 2023, pushing that entry into the next fiscal year’s report, the official explained. An official from the Food Inspection and Safety Division of the MHLW also mentioned that delay, telling WNISR that the ministry’s data was compiled by the date test results were reported rather than when the samples were taken, which could push some entries into a subsequent year.910

A WNISR search of the raw data used by the ministry, contained in the Database of Radioactive Substances in Foods, indicates that 12,311 samples of food produced in Fukushima Prefecture were tested in FY2022.

“Nobody Can Confirm the Data”

Yet even that database is riddled with errors and not reliably updated, said Katsumi Shozugawa, an assistant professor at the University of Tokyo who specializes in environmental analytical chemistry, and has co-authored many studies on contamination in food and the environment after the Fukushima accident.911

Japan has conducted a massive amount of radiation measurements, but the tests have been done by a medley of organizations without an integrated approach or a readily accessible way to figure out what is going on, he said. The Fukushima government as well as organizations in the prefecture that conduct independent food tests have done a good job of monitoring to ensure highly contaminated products do not go to market; other prefectures have not been so thorough, he said. The opaque and uneven testing landscape has left outside observers uneasy and without a way of checking how thorough and reliable Japan’s radiation testing is, he said.

Shozugawa coauthored a study on the safety of the 2020 Tokyo Olympics, prompted by concerns from international athletes worried about exposure risks.912 He concluded the game venues were very safe, but said it took an immense amount of effort to find the data he needed. “It takes an incredible amount of work to grasp the entire picture” of radioactive contamination post Fukushima, said Shozugawa. “This is the biggest problem: nobody can confirm the data.”913

Legal Cases, Compensation

A spate of recent court rulings found TEPCO executives or the government were not to blame for failing to prepare better for risks at Fukushima Daiichi.

Japan’s Supreme Court on 5 March 2025 upheld decisions by lower courts that found two former TEPCO top executives not guilty of criminal negligence leading to the Fukushima Daiichi accidents and the subsequent death of Fukushima evacuees. A third executive, TEPCO chair in March 2011, Tsunehisa Katsumata, died in October 2024, and the charges against him were dismissed at that time.914

The decision by the top court puts a period on the only case to try to assign criminal responsibility for the nuclear disaster. The prosecution had argued that TEPCO could have prevented the accident if it had heeded a long-term tsunami risk assessment and installed proper countermeasures. The Supreme Court agreed with lower courts’ views that the assessment was not considered reliable and the executives could not have forecast the magnitude of the 2011 tsunami based on them. (See also past WNISR editions and Judicial Decisions on Damages and Criminal Liability for the Fukushima Nuclear Accidents in WNISR2021.)

On 6 June 2025, the Tokyo High Court overturned a 2022 lower court decision ordering four former TEPCO executives to pay ¥13 trillion in damages (US$202298.8 billion) to the company. That decision also revolved around the long-term tsunami risk forecasts, with the High Court ruling the forecasts were not seen as pressing, and so the executives should not be held accountable for not acting immediately.915

And the Osaka High Court on 18 December 2024 overruled a 2018 decision by the Kyoto District Court that ordered the government to pay compensation to some Fukushima residents who had evacuated to other prefectures. The Kyoto court had ordered both the government and TEPCO to pay around ¥110 million (~US$2018996,000) in compensation to 110 claimants. The Osaka court revised that to ¥110 million (~US$2024727,000) paid by TEPCO alone to 92 plaintiffs, in the latest decision to shield the Japanese government from responsibility for the accident.916

Meanwhile, the government’s official estimate of the cost of compensation, decontamination, the interim storage facility for contaminated soil and waste, and the decommissioning of Fukushima Daiichi remains unchanged from its latest revision at the end of 2023, at ¥23.4 trillion (US$2023166.6 billion).917 When various administrative costs are added, that estimate rises to ¥26.2 trillion (US$182 billion).918 Other groups have estimated far higher costs. See Estimated Costs of Fukushima Disaster: Official and Independent Assessments in WNISR2021 for various cost estimates.

As of 11 July 2025, the total compensation amount that has already been paid out by TEPCO is ¥11.58 trillion (US$80.3 billion).919 TEPCO’s compensation costs have been rising recently, in part because it is paying for damages suffered by fisheries and other businesses that have seen earnings drop after the company started releasing treated water into the ocean in 2023.920 The releases caused some consumers to shun marine products from the area and a handful of countries like China to ban purchases of Japanese seafood, although some of those restrictions are now coming off.

In March 2025, TEPCO revised its Fourth Comprehensive Special Business Plan, bumping up the cumulative amount of compensation it expects it needs to pay by around ¥180 billion (US$1.2 billion), to ¥13.4 trillion (US$94 billion), which includes the ¥11.5 trillion (US$80 billion) in compensation it had already paid as of March 2025.921

In April 2025, TEPCO released its latest three-year plan for the reserve fund from which it pays its decommissioning expenses, budgeting a total of ¥260.5 billion (US$1.8 billion) for FY2025, including around ¥30 billion (US$208 million) for water management, ¥31 billion (US$215 million) on removing spent fuel, ¥18.7 billion (US$130 million) for the program to retrieve fuel debris, and ¥19.8 billion (US$137 million) on tackling waste. It is budgeting gradual increases for that total over the following two years, with the largest increases in spending on water and waste management.922

1. Overview of Status of the Decommissioning (as of mid-2025)

Sources: compiled by WNISR with TEPCO and METI, Various years

Decommissioning Status Report

Introduction

In mid-2025, 218 nuclear power reactors were closed, corresponding to about 110 GW of permanently retired capacity. This compares with 408 reactors in operation and 33 in Long-Term Outage (LTO). Thus, almost one-third of the reactors ever connected to the grid in the past 70 years have been retired.

Decommissioning nuclear power plants is an important, and often overlooked, element of the nuclear electricity system. Defueling, deconstruction, and dismantling—summarized by the term decommissioning—are the final steps in the operational cycle of a nuclear power plant (excluding waste management and disposal). The process is technically complex and poses major challenges in terms of long-term planning, implementation, and financing. In the first decades of the nuclear age, decommissioning was hardly considered in the reactor design. The costs for decommissioning at the end of a reactor’s the lifetime were usually discounted away, and thus largely ignored. However, as a growing number of nuclear facilities either reach the end of their operational lifetimes or have already been closed, the challenges of reactor decommissioning are increasingly attracting stakeholder and public attention.

Elements of National Decommissioning Policies

When analyzing decommissioning policies, one needs to distinguish between the process (in the sense of the actual implementation) and the financing. To provide a general and globally applicable overview of the progress of nuclear decommissioning, WNISR has been classifying technical decommissioning into three main stages since WNISR2018. This is necessary due to the heterogeneous nature of decommissioning regulations around the world.923 The three stages are defined as follows:

• The warm-up stage comprises the post-operational stage and the dismantling of systems that are not needed for the decommissioning process. In addition, the dismantling of system parts with higher contamination begins, including the defueling of the reactor, which is crucial for any further undertakings and means removing the spent fuel from the reactor core and the spent-fuel pools.
• The hot-zone stage comprises the dismantling activities in the hot zone, i.e., the dismantling of highly contaminated or activated parts, e.g., the reactor pressure vessel (RPV) and its internals (RVI) and the biological shield.
• The ease-off stage comprises the removal of operating systems as well as the decontamination of the buildings. This stage concludes with the end of physical demolition work.

After physical demolition is completed—as far as reported by the owner or the licensing authorities—WNISR classifies decommissioning as “completed”. This does not necessarily reflect the license termination of the site. Generally, a site can be qualified as a brownfield site in which some infrastructure remains for further nuclear or other industrial use, or it can be released from regulatory oversight as a greenfield site for unrestricted use. Different countries require different types of end-states for complete license termination.924 WNISR classifies sites that have been partially released from regulatory oversight but still operate nuclear facilities on site, e.g., interim spent-fuel and waste-storage infrastructure, as brownfield sites.

The technical procedure of physically dismantling nuclear reactors, following the three stages described above, can begin after varying amounts of time following nuclear power plant closure. This depends on the strategy the operator chooses. The options include:

• immediate dismantling, which is characterized by a seamless transition to decommissioning activities after reactor closure
• deferred dismantling, where reactors are placed into Long-Term Enclosure (LTE) for several years to decades to allow for radiation levels to decline before decommissioning begins
• entombment, characterized by LTE (50 years or more) that can sometimes become permanent (e.g., Chornobyl Unit 4)

Most countries have adopted variations of these strategies, although some, like France or Germany, have placed restrictions on which strategy may be applied.925

With respect to financing, five main approaches are observable: public budget, external segregated fund, internal non-segregated fund, internal segregated fund, and surety methods such as guarantees or insurances (for more details, see WNISR2018).926

The goal of this chapter is to provide a global overview of the current state of nuclear decommissioning. However, information on the status and progress at individual sites can be inaccessible, and available data can change over time. This may lead to simplifications regarding the classification of individual sites and reactors into different stages of the decommissioning process. WNISR aims to provide full access to the basis of our assessments and has in the past, when uncertainty arose, communicated individual classification decisions and potential statistical changes and will continue to do so.

Global Overview

Decommissioning Worldwide

As of 1 July 2025, a worldwide total of 218 reactors, corresponding to about 110 GW of capacity, have been closed. Since the reporting period covered in WNISR2024, five additional reactors have been closed: Doel-1 in Belgium (445 MW), the Canadian heavy-water reactors Pickering-1 and -4 (515 MW each), and both reactors at the Maanshan plant in Taiwan (938 MW each).

Of the total number of closed units, 61 percent are located in Europe (106 in Western Europe and 26 in Central and Eastern Europe), 22 percent in North America (49), and 17 percent in Asia (37).

More than three quarters or 171 reactors used one of these three technologies:

• Pressurized Water Reactors (PWRs) with 72 units or 33 percent,
• Boiling Water Reactors (BWRs) with 55 units or 25 percent, and
• Gas-Cooled Reactors (GCRs) with 44 units or 20 percent, the majority (33 units) of which are located in the U.K.

Table 13 provides an overview of the closed reactors worldwide. The table also includes the number of defueled reactors and those that are released from regulatory supervision, i.e., either as ‘full’ greenfield sites or as brownfield. The Decommissioning Status Report, as the WNISR in general, exclusively covers reactors that have generated electricity and were connected to the grid. Thus, it does not cover research reactors that were not connected to the grid.

Decommissioning plays an important and increasing role in nuclear politics, both in the timing and production process, and the financing thereof. The number of facilities that will be affected will increase significantly in the coming years: assuming a 40-year lifetime, a further 93 reactors will close by 2030 (reactors connected to the grid between 1985 and 1990), and an additional 158 will be closed by 2065 (reactors connected 1991–2025). This does not account for the 157 reactors that have already been operating for more than 40 years (connected to the grid before 1985), an additional 33 reactors in Long-term Outage (LTO), and the 63 reactors under construction as of mid-2025. It should be noted, however, that the current general trend is to extend operational lifetimes beyond 40 years, often with the aim of 60 years and in some cases 80 years of operation. (See Figure 21.)

1. Overview of Reactor Decommissioning Worldwide (as of 1 July 2025)

   Country	Closed Reactor	Decommissioning Status
   		Warm-up (of which defueled)		Hot Zone	Ease-off	LTE	Completed (of which released as greenfield sites(a))	Completed Share of Total Closed (share of total closed released as greenfield sites(a))
   U.S.	41	4	(4)(b)	6	5	9	17 (5)	41% (12%)
   Germany	36	10	(4)	10	11	1	4 (3)	11% (8%)
   U.K.	36	19	(12)	11	0	6	0	0%
   Japan	27	25	(4)	1	0	0	1 (1)	4% (4%)
   France	14	3	(2)	3	0	8	0	0%
   Russia	11	6	(5)	0	0	5	0	0%
   Canada	8	3	(1)	0	0	5	0	0%
   Sweden	7	2	(2)	0	5	0	0	0%
   Taiwan	6	6	(0)	0	0	0	0	0%
   Bulgaria	4	4	(4)	0	0	0	0	0%
   Italy	4	3	(3)	1	0	0	0	0%
   Belgium	4	3	(0)	0	1	0	0	0%
   Ukraine	4	0	(0)	0	0	4	0	0%
   Spain	3	1	(0)	0	0	1	1 (0)	33%
   Slovakia	3	1	(1)	0	2	0	0	0%
   Lithuania	2	2	(2)	0	0	0	0	0%
   South Korea	2	2	(0)	0	0	0	0	0%
   Armenia	1	0	(0)	0	0	1	0	0%
   Kazakhstan	1	0	(0)	0	0	1	0	0%
   Netherlands	1	0	(0)	0	0	1	0	0%
   Switzerland	1	0	(0)	1	0	0	0	0%
   India	1	1	(1)	0	0	0	0	0%
   Pakistan	1	1	(0)	0	0	0	0	0%
   Total	218	96	(45)	33	24	42	23 (9)	11% (4%)

Sources: Various, compiled by WNISR, 2025

Notes:

LTE: Longt-Term Enclosure.

(a) As of WNISR2024, the “released” category only contains “greenfield released” reactors, whereas it also contained “brownfield released” reactors in previous editions (see definitions in the introduction of the chapter).

(b) Three reactors in the U.S. are currently undergoing preparations or are being considered for a restart. These are Palisades, Three Mile Island-1 (now called Christopher M. Crane Clean Energy Center or CCEC), and Duane-Arnold-1. Until the reactors are restarted and this is officially confirmed, WNISR classifies them in their respective last known decommissioning phases (warm-up for Palisades and LTE for the others).

(c) In WNISR2024, the Trino reactor in Italy was erroneously placed into the hot-zone stage and is therefore listed in the warm-up stage in this edition (see section on Italy below).

Overview of Reactors with Completed Decommissioning

As of mid-2025, 195 units globally are awaiting or in various stages of decommissioning, five more than in WNISR2024. This number includes three units in the U.S. that might restart.

Of the 23 decommissioned reactors, 21 have been released from regulatory oversight, nine of those as greenfield sites (see Figure 52). All other sites were released as brownfield sites, and apart from the Shoreham reactor in the U.S., all currently host interim waste storage facilities.

The average duration of the decommissioning process, independent of the chosen strategy, is around 20 years, with a very high variance: the minimum being six years for the 22-MW Elk River plant and the maximum 45 years for the 63-MW reactor at Humboldt Bay, both in the U.S.

Only four countries amongst the 23 with closed power reactors have completed the technical decommissioning process of at least one reactor: the U.S. (17 units), Germany (4), Japan (1), and Spain (1). Some of the reactors that were most rapidly decommissioned are located in the U.S. In Germany, the HDR (Heißdampfreaktor, a superheated-steam reactor) Großwelzheim was only on the grid for one year, but decommissioning lasted well over 20 years. The German Würgassen reactor has de facto completed the technical decommissioning process but, legally, cannot be released from regulatory control as the buildings are being used for interim storage of wastes.927 In Japan, the only reactor to be decommissioned was a small 10-MW demonstration plant (JPDR), whereas none of the large commercial reactors have been decommissioned yet.928 It took over 17 years to decommission the José Cabrera reactor, Spain’s first PWR that operated from 1968 to 2006.929 Figure 53 provides the timelines of the 23 reactors that have completed the decommissioning process.

1. Overview of Completed Reactor Decommissioning Projects

Sources: Various, compiled by WNISR, 2025

Note: As of WNISR2024, the “released” category only mentions “greenfield released” reactors, whereas it also contained “brownfield released” reactors in previous editions (see definitions in the introduction of the chapter). The LaCrosse and Shoreham reactors had in previous WNISR editions been classified as “greenfield released”. However, the site at LaCrosse still contains an interim spent-fuel storage facility and is thus reclassified as a brownfield site, and while the Shoreham reactor was released from regulatory oversight, it was never entirely dismantled, and most buildings have remained intact.

Overview of Ongoing Reactor Decommissioning

This section contains a brief overview of the decommissioning status in the countries that have not been analyzed in more detail in the Country Case Studies of this WNISR edition.

Armenia

Following a partnership agreement with the European Union, the Armenian Metsamor nuclear power plant is to be completely closed as soon as possible due to safety concerns because the plant “cannot be upgraded to fully meet internationally accepted safety standards.”930 Both units were closed in 1989 after an earthquake, but Unit 2 was restarted in November 1995 and still operates to this day, with ongoing lifetime extension plans aiming for it to run until 2036 (see section on Armenia).931 At Unit 1, the decommissioning status remains unclear. While the Armenian government adopted a decommissioning strategy in November 2007, at the time aiming for a deferred dismantling strategy to conclude in 2048,932 and an EU-funded project “to support the Decommissioning Planning and Licensing for general decommissioning activities” ran from 2013 to 2018,933 there still seems to be no officially selected strategy. International Atomic Energy Agency (IAEA) data from 2024 shows the strategy as “other,”934 but data from 2019 suggest a deferred dismantling strategy with ongoing defueling activities.935 Since there have been no publicly communicated updates regarding the progress at Unit 1, WNISR considers the reactor to be in LTE.

Belgium

In Belgium, the prototype 10-MW BR-3 reactor in Mol, which was closed in 1987, is currently undergoing decommissioning. The reactor is in the ease-off stage936 and is used as a lead-and-learn site for future decommissioning projects.937 In November 2024, the dismantling of the biological shield was completed.938 Apart from Doel-4 and Tihange-3, whose state-funded lifetime extensions to 2035 were approved by the European Commission in February 2025,939 all remaining power reactors are scheduled to be closed by the end of 2025 (see section on Belgium). Tihange-1 and Doel-2 are to close by October and December 2025, respectively,940 and three reactors (Doel-1, Doel-3, and Tihange-2) have already been closed. WNISR considers Doel-3 and Tihange-2 to be in the warm-up stage, with all the fuel removed from the two reactor vessels and currently stored in the spent-fuel pool.941 Dismantling of reactor internals necessitates a separate decommissioning license to be granted; this phase is expected to begin for both reactors in 2026.942 Doel-1 was closed on 14 February 2025 and is considered in the warm-up stage. However, decommissioning activities will only begin once Doel-2 has also been closed because the reactors are considered “Siamese twins” as they “share a control room, machine room, and also a lot of common safety systems.”943

Bulgaria

Decommissioning of all four closed reactors at the Bulgarian Kozloduy plant is partly funded by the Kozloduy International Decommissioning Support Fund managed by the European Bank for Reconstruction and Development. The fund had raised more than €1.2 billion (US$20231.3 billion) by the end of 2023 and supports the plan to decommission all four units to a brownfield state by 2030.944 In its program review, the E.U. states that “the timely achievement of the objectives for 2021-2027 is […] at risk.”945 The most recent report on the whole program was published in April 2024 and refers to activities conducted in 2022,946 but the European Commission tracks the project on a dedicated website that is more up to date.947 Primary decontamination work has been completed at all four units, removing up to 99 percent of activated material. While the official completion target remains at 2030, the operator is likely to postpone it to 2032. In late 2024, the Bulgarian regulator renewed the decommissioning license for Units 1 and 2 until 2034.948 Until information on hot zone dismantling becomes available, WNISR considers all four units to be in the warm-up stage. According to Bulgaria’s latest report to the Joint Convention on the Safety of Spent Fuel Management and on the Safety of Radioactive Waste Management, “the entire amount of nuclear fuel has been removed from the reactors;”949 all four reactors are thus considered defueled.

Canada

As of mid-2025, Canada had eight closed CANDU (Canada Deuterium Uranium) and one Heavy-Water Moderated Boiling Light-Water Reactor (HW BLWR) to decommission. The most recent closures were those of the Pickering-1 and -4 reactors, which ceased operations in October and December 2024, respectively. So far, no commercial reactor has been fully decommissioned. The 250-MW Gentilly-1 prototype HW BLWR reactor was closed in 1977 after only 183 full-power-equivalent days of operation and was subsequently defueled in 1984 and placed into LTE in 1986, where it has remained since. The current license will terminate in 2034.950 The prototype CANDU reactor at Douglas Point operated from 1967 to 1984 and was subsequently placed into LTE. Active dismantling began in 2021 with plans to complete decommissioning by 2070, though mention of that deadline has since been removed from the official webpage.951 In March 2025, the demolition of the administrative building and the ancillary facilities was completed; the project is now moving on to demolishing the turbine hall, before beginning decommissioning work of the reactor internals and respective buildings, which is to be completed after 2030.952

The other six closed reactors are former commercially operated CANDU reactors. The Gentilly-2 reactor was closed in 2012 and defueled in 2013. It was subsequently placed into LTE with dismantling to begin in 2057. The site is planned to be remediated by 2064, while an environmental follow-up is scheduled to conclude by 2074.953 In 2023, owner-operator Hydro-Québec had toyed with potentially restarting the reactor,954 but any plans seem to have been scrapped with the application in April 2025 for a decommissioning license renewal, including for “the early dismantling of buildings and structures containing little or no radioactive contamination.”955 Two reactors of Pickering-A, Units 2 and 3, were already closed in 1997 and have been defueled, while Units 1 and 4 were closed in September and December 2024, respectively.956 Once defueling is completed, these units will join Units 2 and 3 in LTE for concrete planning for dismantling to begin in a yearly staggered approach from 2050 onwards, with practical work expected to begin in 2051. LTE is scheduled to start in 2029. Until then, Units 1 and 4 are considered to be in the warm-up stage. The four still operational reactors of the Pickering-B plant are scheduled to follow this approach from 2054. Pickering-A is to be released as a brownfield site by 2063.957

France

The closed reactor fleet in France is diverse compared to the largely standardized operational PWR fleet. In total, 14 reactors (8 GCRs, 3 PWRs, 1 HWGCR, 2 FBRs), corresponding to approximately 5.5 GW, have been closed. Decommissioning progress is slow in France (see previous WNISR editions). All GCRs, the six UNGG-type (Uranium Naturel Graphite Gaz) reactors, Bugey-1, Chinon A-1, A-2, A-3, and St. Laurent A-1 & A-2, as well as the G-2 and G-3 reactors at Marcoule are considered as being in LTE by WNISR.958 For its six UNGG-GCR reactors, EDF initially adopted an immediate dismantling strategy in 2001 to flood the reactor vessel with water, followed by plans to perform decommissioning procedures underwater.959 Following a regulatory change, this strategy was changed to in-air dismantling in 2016, prompting the need for new decommissioning licenses for each reactor. These were submitted in December 2022 and are expected to be granted in 2026, at the earliest. Based on the revised plans, the first dismantling work of reactor internals and graphite blocks will begin at Chinon A-2 in 2044 after the first opening of the reactor casing in 2034. All other UNGG reactors will remain in LTE until at least 2056. Completion of decommissioning is planned for “between 2063 and 2093, depending on the reactors.”960

The EL-4 reactor at Brennilis (Monts d’Arrée) was closed in 1985 and is in the hot-zone stage. With additional regulatory approval, dismantling of the reactor block and demolition of the containment building began in November 2024. Site remediation is expected for 2041.961

The 305-MW PWR reactor at Chooz-A was closed in 1991 and has been undergoing decommissioning work since 2007. In 2023 and 2024, preparatory work for reactor internal dismantling was ongoing, including pool drainage and primary circuit segmentation. Reactor vessel dismantling is scheduled to begin in 2027 (two years later than last year’s estimate, see Decommissioning Status Report—Country Case Studies: France in WNISR2024).962

The two PWRs at Fessenheim were closed in 2020. EDF planned for a five-year preparatory phase until the decommissioning license is obtained, which was expected to be issued in mid-2025 and take effect in early 2026. In the meantime, both reactors have been defueled, full system decontamination has been completed, and dismantling work in the turbine hall is ongoing.963

The FBR Super-phenix at Creys-Malville, closed in 1996, has been undergoing decommissioning since 2006. Currently, reactor vessel internals are being dismantled, with the neutron shield support (the first part of the vessel internals) having been removed in 2024 and dismantling continuing in the reactor pit. The current decommissioning finalization target is 2034.964 Decommissioning at the FBR Phénix at Marcoule continued with defueling activities in 2024, after which further dismantling shall begin. The reactor’s sodium will be transferred to the storage facility of the NOAH965 sodium treatment plant, scheduled to become operational in 2028.966

Since 2020, there have been no publicly announced updates regarding dismantling progress at the remaining GCR plants, G-2 and G-3, formerly operated by CEA at Marcoule.967 Thus, WNISR considers these reactors to be in LTE.

India

Rajasthan-1 in India—in LTO (Long-Term Outage) status since 2004 and considered as closed by WNISR since 2014—has apparently not moved beyond complete defueling and is reportedly “maintained under dry preservation.”968 WNISR considers this PHWR to be in the warm-up phase.

Italy

Since 1988, Italian nuclear power plants have not produced any electricity, and the last two reactors were officially closed in 1990. Since then, decommissioning at all four facilities has been underway. The work is conducted by Italian company Società Gestione Impianti Nucleari SpA (Sogin). The smallest reactor, Garigliano, a 150-MW BWR, was officially closed in 1982 and is considered in the hot-zone stage.969 According to the 2024 report of Italy to the Joint Convention on the Safety of Spent Fuel Management and on the Safety of Radioactive Waste Management, brownfield is to be achieved in 2028 and greenfield in 2042970 (see Table 14); however, the Italian project website claims a target year of 2031.971 At the Enrico Fermi (Trino) plant, decommissioning work has been ongoing since 1999. Tenders for dismantling reactor internals are being prepared.972 Official reports note that most of the work that happened at Trino in 2024 were preliminary activities to the actual dismantling, such as the design, commissioning and licensing preparations for the project, and that in 2024, “no dismantling operations with significant impacts were carried out,”973 which stands in contrast to a February 2024 news report that claims that reactor internals dismantling has begun.974 Until more information becomes available, WNISR classifies Trino in the warm-up stage. At Italy’s largest reactor, the 860-MW BWR Caorso, Sogin reported in November 2024 that it had begun dismantling external components of the reactor building.975 Since there have been no updates on any hot zone work, WNISR considers Caorso to be in the warm-up stage. The Latina GCR is awaiting the operation of Italy’s planned waste repository for reactor dismantling to begin, currently envisioned for 2039.976 Until then, auxiliary buildings are being demolished during the so-called “Phase 1” of decommissioning.977 The reactor core is defueled.978

1. Target Dates for Decommissioning Milestones in Italy

Sources: Reactor data compiled by WNISR with IAEA-PRIS, 2025; and Decommissioning data from Italy’s 7th National Report to the Joint Convention on the Safety of Spent Fuel Management and on the Safety of Radioactive Waste Management, 2024

Japan

As of mid-2025, 27 reactors (17.1 GW) were permanently disconnected from the Japanese grid. Almost all currently closed reactors remain in the warm-up stage as of now, except Hamaoka-2 that entered the hot-zone stage. The country, one of the early adopters of nuclear power, has not completed the decommissioning of a single commercial reactor. The only completed decommissioning project is the small 12-MW research reactor, Japan Power Demonstration Reactor (JPDR), which was released as a greenfield site in October 2002.979

Cleanup at the Fukushima Daiichi plant is slowly advancing as fuel is being removed from the reactors. Fuel removal from the spent-fuel pools has been completed at Units 3 and 4, and this step is expected to begin for Units 1 and 2—which both suffered meltdowns along with Unit 3—in FY2027–2028 and FY2026, respectively. Spent-fuel assemblies from Unit 6 have already been retrieved, although some fresh fuel is still stored in the pool and reactor building, while the defueling of Unit 5 is to begin in July 2025.980 The aim is for all spent-fuel pools to be empty by 2031.981 Fuel debris removal however is in its preliminary stages. TEPCO retrieved a first sample in late 2024—0.9 grams in total—and is still to determine how to remove and safely store the estimated remaining 880 tons. Actual dismantling is yet to begin, and the official target for completion of decommissioning is 2051. See Fukushima Status Report for details.

At the other plant in the Fukushima prefecture, Fukushima Daini, all four BWRs are in the process of being defueled,982 with decommissioning scheduled to conclude in FY2064.983

Dismantling work on Hamaoka-1 and -2 began in late 2009, and the planned finalization for both BWRs is FY2042.984 In March 2025, the dismantling of the Reactor Pressure Vessel (RPV) began at Hamaoka-2 and will be followed by that of Unit 1; the process is expected to take about six years per unit.985 Thus, Unit 2 is in the hot-zone stage, while Unit 1 is still in warm-up. The decommissioning of the Magnox reactor Tokai-1 began in 2001. The dismantling of auxiliary components and buildings has begun, and in some cases, such as for the turbines, it has been completed. Decommissioning activities are scheduled to end in 2035.986 According to a news report, the completion date was initially set for FY2017 and has already been postponed four times.987 At all other closed reactors, no meaningful progress has been reported since the Decommissioning Status Report in WNISR2024. Decommissioning completion dates for all Japanese reactors range from FY2030 to FY2064, with most planned in the late 2040s and early 2050s.988

Kazakhstan

Decommissioning has been underway since 1998 at Aktau BN-350, a sodium-cooled fast reactor in Kazakhstan. The reactor is currently being transferred into LTE status, where it will remain for 50 years, after which dismantling shall begin.989 A publication by the Institute of Atomic Energy of Kazakhstan’s National Nuclear Center in early July 2025 stated that

recently, the work on decommissioning the BN-350 reactor has significantly decreased. The main reason is associated with insufficient financing of this project. Consequently, activities related to the completion of the second, third, and fourth stages [of decommissioning] were almost suspended. At the same time, the reactor facility is maintained in a safe state to prevent dangerous situations for people and the environment.990

According to a July 2024 report, part of the KZT125 billion (US$2020303 million) decommissioning budget, funded through an electricity tariff on local residents, is proposed to be used for the construction of a museum whose building would “replicate the main building of the BN-350.” As of July 2024, KZT16 billion (US$202434 million) had reportedly been paid for design work alone.991

Lithuania

In Lithuania, the two Ignalina Soviet-designed Chornobyl-type RBMK units with 1185 MW capacity each have been closed since 2004 and 2009, respectively. Both reactors and their spent-fuel pools have been defueled. In 2024, dismantling activities advanced, with 24 demolished buildings. By the end of the year, 44 percent of the total planned volume of equipment and structures had been dismantled.992 Hot-zone works, i.e., reactor internals dismantling, are scheduled to begin in 2028, with brownfield remediation of the site planned for 2038.993 (See WNISR2019 for details on decommissioning in Lithuania.)

Netherlands

In the Netherlands, the 55-MW reactor Dodewaard was defueled in 2003 and placed into LTE in 2005 to begin actual dismantling activities from around 2045 onwards, with the objective to return the site to greenfield status.994 With the former operator Gemeenschappelijke Kernenergiecentrale Nederland (GKN), owned by Nederlands Elektriciteit Administratiekantoor (NEA), having mismanaged dedicated decommissioning funds (see earlier WNISR editions), the ownership of the plant was transferred to the Dutch state-owned radioactive waste management company COVRA for the symbolic fee of €1 in December 2024.995

Pakistan

Pakistan closed its first reactor Karachi Nuclear Power Plant-1 (KANUPP-1) in 2021, with a decommissioning license granted in 2022, marking the beginning of preparatory work to place the reactor into LTE for 20–30 years, after which dismantling to a brownfield status is planned.996 With LTE preparations ongoing—expected to last about 15 years—and no published changes or updates since last year’s edition, WNISR considers the reactor to be in the warm-up stage, to be later transferred to an LTE status.

Russia

As of mid-2025, Russia accounts for 11 closed reactors with a combined capacity of 4.9 GW, consisting of two different reactor types: eight first-generation Light-Water Gas-cooled Reactors (LWGR)—among them four RBMK Chornobyl-type reactors—and three Soviet-designed PWRs. Kursk-2 was the latest reactor to close in Russia in January 2024.997 Kursk-1 and -2 and Leningrad-1 and -2 are the four closed RBMK reactors that pose substantial challenges for decommissioning because of the large volume of irradiated graphite stacks. While robotic dismantling techniques were being tested before Russia invaded Ukraine, the research progress remains uncertain.998 The two closed reactors at the Leningrad plant shall be turned into pilot sites for RBMK dismantling.999 Both reactors were defueled as of December 2024,1000 and they are awaiting the dedicated decommissioning licenses to be approved, which is expected sometime during 2025, before actual dismantling can begin. In the meantime, additional onsite waste storage and treatment facilities are to be built by 2026.1001 All four RBMKs are considered in warm-up stage. At the Novovoronezh plant, the initial deferred dismantling strategy was changed to an immediate dismantling approach in 2017 for Units 1 and 2 as a pilot project for Soviet/Russian PWR dismantling.1002 In 2024, the project apparently moved on to active dismantling and decontamination of material and components; WNISR thus considers the units to be in the warm-up stage.1003 Considering the lengthy estimated decommissioning duration of 50 years and unclear decommissioning strategies at the other sites,1004 WNISR considers all other Russian reactors to be in LTE as long as there is no documented evidence of decommissioning progress. (See WNISR2019 for details on decommissioning in Russia.)

Slovakia

Slovakia’s decommissioning efforts are advancing, with both 408-MW VVER-230 reactors Bohunice-1 and -2 in the ease-off stage.1005 The decommissioning company Jadrová a vyraďovacia spoločnosť (JAVYS) estimates total costs for both reactors at €1.22 billion (US$1.4 billion), but project performance is not updated regularly, complicating an assessment of the accuracy of this estimate and of the project timeline.1006

Bohunice A1, a 93-MW heavy water GCR-type reactor, began decommissioning in 1999. All fuel has been removed from the site since 2009.1007 Currently, the reactor is in the fourth of five decommissioning phases, which was planned to conclude in 2024.1008 Decommissioning will continue with the dismantling of reactor internals in the fifth phase. As of the time of writing in mid-2025, there have been no official updates on the completion of the fourth phase,1009 and the licensing process for the fifth stage is yet to be completed. The whole project is expected to be completed by 2033.1010

South Korea

As of mid-2025, two commercial reactors were closed in South Korea. The first reactor, South Korea’s oldest unit, Kori-1, a 576-MW PWR, was closed in 2017. At the time, defueling was envisioned for end-2025 and the completion of plant dismantling by 2032.1011 Operator Korea Hydro & Nuclear Power (KHNP) applied for a decommissioning license in 2021. As of early 2025, the license was expected to be granted “in the first half of [2025].”1012 The regulator issued its approval in late June 2025. Cost of decommissioning was estimated at KRW1.7 trillion (US$1.2 billion), for a 12-year implementation duration.1013 Meanwhile, preparatory work, such as chemical decontamination, has already begun.1014 Wolsong-1, a 661-MW Pressurized Heavy-Water Reactor (PHWR), ceased generating power in May 2017 and was officially closed in December 2019.1015 KHNP submitted the final decommissioning plan for the Wolsong-1 reactor in June 2024 and is awaiting final approval by the end of 2026 at the earliest.1016 Until then, preparatory work is underway, with the current schedule aiming for site remediation in 2034.1017

Spain

Current Spanish legislation plans to close all seven still operating units by 2035. In February 2025, a non-binding proposal to extend reactor lifetimes passed the Spanish Congress,1018 and the Iberian blackout in late April 2025 has rekindled the debate on the Spanish phaseout (see section on Spain).1019 Regardless, decommissioning at the three closed reactors is continuing.

Since WNISR2024, the José Cabrera (Zorita) reactor, which was closed in 2006, is considered fully decommissioned. Currently, final remediation work is underway with the site reportedly planned to be released as a greenfield site by “the end of 2027 or early 2028.”1020

The GCR Vandellos-1 is scheduled to remain in LTE until 2030, after which reactor internals dismantling will begin.1021 Earlier versions of the program envisioned the start of these activities and the end of LTE in 2028.1022 As of 2024, Empresa Nacional de Residuos Radiactivos S.A (Enresa) indicated that the last phase of dismantling would last about 15 years, and the site would be released of regulatory oversight after a maximum of ten years following the end of works, meaning around 2055 at the latest.1023

The 446-MW BWR Santa Maria de Garona (Garona-1) that last produced electricity in 2012, suspended operations in 2013 and was officially closed in 2017. The first decommissioning phase, planned from 2023 to 2026, is underway. It includes defueling procedures and dismantling of the first auxiliary components, such as the turbines.1024 According to a news report from November 2024, this phase is delayed by two years.1025 It remains unclear whether this delay will impact the target of site remediation by 2033, as Enresa’s website has not been updated accordingly as of mid-2025. (See WNISR2019 for details on the Spanish decommissioning process.)

Sweden

In Sweden, the closed reactor fleet comprises seven reactors, corresponding to 4.1 GW, and consists of five BWRs, one PWR, and one PHWR. With the completion of RPV dismantling in late 2023,1026 and the removal of the steam generators from the reactor building in January 2025,1027 the Ågesta PHWR, which was closed in 1974, has moved on to the ease-off stage. The plan is for the project to conclude in 2027.1028

Reactors at Barsebäck and Oskarshamn (two BWRs each) have moved to the ease-off stage with the completion of RPV segmentation by a consortium of German companies Nukem and Uniper Nuclear Services in December 2024.1029 Remaining dismantling and demolition is expected to be completed at both sites by 2028.1030 The plan is to release both plants as brownfield sites subsequently.1031

Ringhals-1 and -2 were closed in 2020 and 2019, respectively. Defueling was completed in May 2022, when all spent fuel was transported to the Swedish interim waste storage facility CLAB.1032 In late 2023, Nuvia, a subsidiary of French construction company Vinci, was awarded the contract to dismantle both plants. The projects are to be completed by 2031, and reactor internals will be dismantled by Westinghouse following a 2021 tender award.1033 First dismantling work began in 2024, seemingly limited to external components; thus, both reactors remain in the warm-up stage until further notice.1034 In June 2025, three steam generators were removed from Unit 2.1035

Switzerland

Switzerland has some decommissioning experience, having completed technical decommissioning of the research reactor at Lucens in 2004.1036 Decommissioning of the commercial reactor at Mühleberg began shortly after its closure in 2019. The site has been defueled since September 2023.1037 In 2024, the control rods in the reactor core were dismantled, and work was ongoing to empty the spent-fuel pond.1038 The reactor is therefore considered to be in the hot-zone stage. Dismantling of nuclear components is expected to be completed in 2031, with the site being released for further use by 2034.1039 Per latest announcements, operator BKW is considering the construction of a battery storage facility or a “low-emission gas power plant.”1040 An application to repurpose the site must be submitted before the end of 2027, and as of late 2024, BKW was still reviewing its options.1041

Ukraine

In Ukraine, the three reactors Chornobyl 1–3 completed defueling activities in 2016,1042 and following the chosen deferred dismantling strategy, they are to be placed into LTE from approximately 2028 to 2045. They will then be dismantled for regulatory release in 2065.1043 At the damaged Unit 4, the so-called New Safe Confinement was completed in 2016.1044 In December 2023, the dismantling license to stabilize critical parts of the original 1986 sarcophagus—“whose collapse probability is very high”—and to subsequently demolish it was extended for another six years to 2029. The original plan had been to complete these tasks by 2023.1045 On 14 February 2025, the outer wall of the confinement was hit by a drone strike, resulting in several fires breaking out.1046 In March 2025, the situation was downgraded to a “controlled situation” from the initial “emergency” category, but several more drone sightings had been reported.1047

United Kingdom

In the United Kingdom (U.K.), 36 reactors, corresponding to 7.75 GW, are awaiting decommissioning. Over the past few years, the organization of nuclear decommissioning has been reintegrated into the responsibility of the Nuclear Decommissioning Authority (NDA), which will decommission all British reactors, apart from the PWRs Sizewell B and Hinkley Point C, via its various so-called “Site License Companies”.1048

As of mid-2025, 19 reactors were in the warm-up stage (of which 12 were defueled), 11 in the hot-zone stage, and six reactors remained in LTE. The U.K. is facing the challenge of decommissioning its so-called legacy fleet, consisting mostly of Magnox GCRs, whose decommissioning had been neglected for decades. Refer to earlier WNISR editions for more details on decommissioning in the United Kingdom. The following paragraphs only provide examples of status and ongoing work at select sites.

The British Advanced Gas-cooled Reactors (AGR) are operated by French state-owned utility Électricité de France (EDF)’s subsidiary EDF Energy. Eight reactors remain operational but are expected to close by 2030. After defueling, all AGRs will be transferred to NDA stewardship for further dismantling.1049 Three AGR plants are currently undergoing defueling. At Dungeness B, defueling began in 2021, but no completion date is publicly available.1050 In contrast, defueling at Hunterston B, where two reactors were closed in 2021 and 2022, respectively, was announced as completed in April 2025.1051 The first handover to the NDA is planned for 2026, after which the site will be transferred into an LTE state for 70 years.1052 At Hinkley Point B, which is housing two reactors that were both closed in 2022, the formal handover to the NDA is expected for 2026, after which an equally long LTE period shall begin in 2039.1053

Decommissioning of the Magnox fleet is advancing slowly. With only the four Calder Hall units and both reactors at Bradwell remaining in LTE, most sites have moved on to active dismantling, with physical demolition target dates ranging from 2050 to 2080.1054 The NDA’s current strategy draft, open for consultation and scheduled for release in 2026, plans for all buildings to be decommissioned, demolished, or reused by 2089.1055 Whether this target stays in the final version of the strategy remains to be seen. Since WNISR2024, only the two 195-MW reactors at Trawsfynydd have made significant advancements, moving from the warm-up into the hot-zone stage. Dismantling of the fast breeders PFR and DFR at Dounreay is expected to begin in the coming years. The Winfrith prototype AGR shall be fully decommissioned by 2036.1056

Since the restructuring of nuclear decommissioning in the U.K., organizational improvements have been acknowledged by the government’s watchdog agency National Audit Office; however, especially at Sellafield, one of the most complex nuclear sites in Europe, remaining inefficiencies have led to a 6–13-year delay of the clean-up of legacy ponds and silos, and the NDA is still failing to provide “value for money”, according to a report to Parliament in October 2024.1057 WNISR will continue to monitor decommissioning progress in the U.K.

United States

The United States (U.S.) has not only the largest fleet of operating and closed reactors, but also the highest number of fully decommissioned units, representing nearly three-quarters of the global total. In the U.S., as of mid-2025, 41 reactors (20 GW) were closed.1058 Of these, 17 units or 7.1 GW have been fully decommissioned. Currently, decommissioning work is ongoing at 15 units: four units are in the warm-up stage (all defueled), six are in the hot-zone stage, five in the ease-off stage, while nine remain in LTE. Refer to previous WNISR editions for more details on individual decommissioning projects.

The Dresden-1 BWR was not reopened after a 1978 routine inspection because post-TMI (nuclear accident which occurred at the Three Mile Island plant in the U.S. in 1979) safety upgrades were deemed too costly.1059 The reactor has been in LTE since 2007. According to U.S. regulations, decommissioning must be completed within 60 years of reactor closure, i.e., 2038 for Dresden-1. In March 2024, however, operator Constellation requested an extension of this deadline from the U.S. Nuclear Regulatory Commission (NRC) because of plans to extend the operational lifetimes of the still-running reactors Dresden-2 and -3 by 20 years to 2049 and 2051, respectively, and the hopes of dismantling all three plants simultaneously.1060 As of mid-2025, there was no official ruling on this application.1061 At the Peach Bottom-1 High Temperature Gas-Cooled Reactor (HTGR), closed since 1974 and in LTE since 1978, Constellation applied for a similar extension in 2023 to dismantle the reactor jointly with the still-operational Units 2 and 3. Here, too, additional lifetime extensions have been applied for, potentially extending operations beyond 2050.1062 The application for Peach Bottom-1 is yet to be approved by the NRC.1063

The 24 MW BWR GE Vallecitos was closed in 1963 and was subsequently placed in LTE. No major dismantlement activities were carried out before 2007–2008.1064 In 2023, it was announced that private decommissioning company NorthStar would acquire ownership of the plant pending regulatory approval.1065 In November 2023, the removal of the reactor vessel was completed.1066 The agreement to transfer ownership of the entire complex of Vallecitos Nuclear Center—where several research reactors are hosted—from GE Vernova and GE-Hitachi to NorthStar was closed in March 2025.1067 WNISR considers the reactor to be in the ease-off stage. At the Pilgrim-1 reactor, currently being decommissioned by Holtec International, disputes between Holtec and local and state authorities regarding the discharge of radioactive wastewater into Cape Cod Bay, Massachusetts, are ongoing.1068 According to Holtec, the dispute and its regulatory fallout have already delayed decommissioning by years and are prolonging the process.1069

In the U.S., several closed reactors are being considered for restart. The 601-MW BWR Duane Arnold-1 was closed in 2020 after equipment, including the cooling towers, was damaged by a derecho, a type of severe thunderstorm event.1070 The reactor was moved into LTE and was scheduled to be decommissioned by 2080.1071 In March 2025, however, licensee NextEra Energy engaged in talks with the NRC regarding a potential restart of the reactor by 2028.1072

Unit 1 at Three Mile Island was closed in 2019. In September 2024, operator Constellation signed a 20-year PPA with Microsoft to restart the 837-MW PWR. Current plans envision operations to begin in 2028 and for the plant to run to at least 2054.1073 The reactor’s name was officially changed to Christopher M. Crane Clean Energy Center (CCEC) in May 2025.1074

In early 2023, Holtec announced its plans to restart the 805-MW PWR at Palisades (closed since May 2022) and received substantial financial support from the U.S. Department of Energy and the state of Michigan. While Holtec plans for the reactor to reopen in late 2025, several indications on steam generator pipes, found during a September 2024 NRC inspection, necessitate substantial refurbishment. The NRC might not approve Holtec’s plans of “sleeving” the tubes.1075 Alternatively, the steam generators would have to be replaced, endangering the ambitious restart plans.1076 Holtec expects the sleeves will extend the steam generators’ operating life by 30 years and is sticking to the restart of Palisades “in the fourth quarter of 2025.”1077

Until a reactor is officially restarted, WNISR considers them to be in their respective decommissioning stage (see United States Focus for further details).

Decommissioning in Selected Countries

This section provides a more detailed update on decommissioning developments in the two most recent nuclear phaseout countries, Germany and Taiwan.

Country Case Studies

Germany

Despite Germany’s closure of its last three reactors, Emsland, Isar-2, and Neckarwestheim-2, in April 2023 after their planned closure dates were postponed by about three months in the wake of the energy crisis caused by Russia’s invasion of Ukraine in 2022 (see Germany Focus in WNISR2023), the debate on Germany’s nuclear phaseout was reignited in the run-up to the federal election held in February 2025. Most notably, the conservative Christian Democrats (CDU) together with their sister-party CSU (the conservative equivalent in the federal state of Bavaria) criticized the closure as an “ideological mistake” made by the so-called “Traffic Light” (Ampel) coalition (formed by Social Democrats (SPD, red), the Green Party, and the Liberal Democratic Party FDP (yellow).1078 In April 2024, CDU politicians suggested halting the ongoing decommissioning work at the three latest closed reactors. They proposed corresponding legislation in the German Parliament. The initiative was rejected against the votes of CDU, CSU, and the far-right AfD (Alternative für Deutschland).1079 Nonetheless, in the run-up to the election, these political parties defended the proposal to halt nuclear decommissioning regardless of the stage of decommissioning and despite all three nuclear power plant operators rejecting plans to reopen closed plants and the nuclear option in general. Moreover, advocating for reintegrating nuclear power into Germany’s energy supply, both CDU and CSU emphasized on SMRs, so-called “advanced reactors”, and nuclear fusion.

After the election, designated Chancellor Friedrich Merz (CDU) reiterated the supposed necessity of halting decommissioning work.1080 Although the coalition agreement between CDU, CSU, and SPD to form a government does not mention the terms “nuclear power”, “decommissioning”, or “nuclear waste storage” (or any synonyms except for “nuclear fusion”, an entirely different topic),1081 and the propagated announcement of reactivating closed reactors seems to have evaporated,1082 it is useful to assess whether restarting German reactors would have been technically feasible at all or whether the announcements were made merely for election campaigning reasons.

German nuclear legislation requires “immediate dismantling” of closed nuclear power plants.1083 Consequently, decommissioning work began at the last three operating plants once the respective decommissioning licenses had been obtained.

One of the major steps of reactor decommissioning is the decontamination of the primary loop. Decontamination measures can be taken at operating nuclear power plants to reduce the activation content of components and reduce the radiation exposure to personnel. During in-operation decontamination, less aggressive methods are applied to ensure that components are not damaged and remain safe for continued operation. However, when a plant has been closed, so-called Full System Decontamination (FSD) takes place to substantially reduce activation and further reduce radiation exposure to workers. This includes the decontamination of the primary loop.1084 Several decontamination techniques exist,1085 but the most commonly used method involves injecting diluted acids into the system to dissolve deposited radioactive materials that have accumulated in the pipes and also dissolve some of the pipes’ material to remove activation that has dissipated into the pipes themselves.1086 Consequently, once FSD has happened at a closed reactor, e.g., when the primary loop has been decontaminated, operations cannot be restarted without replacing the decontaminated components, which—as the head of EnBW (operator of Neckarwestheim-2), Georg Stamatelopoulos, also pointed out—would require substantial investment and several years to execute.1087 Table 15 provides an overview of the decontamination status at the nine reactors closed since 2011.1088

Since 2018, WNISR analysis has shown that decommissioning work is advancing at all German reactors. Currently, with one reactor in LTE, decommissioning work is being conducted at 31 reactors:

• Ten reactors are in the warm-up stage: Biblis-A (defueled), Brokdorf, Emsland, Grohnde, Gundremmingen-B (defueled) & -C, Isar-2, Krümmel (defueled), Lingen (defueled), and Neckarwestheim-2.
• Ten reactors are in the hot-zone stage: AVR Jülich, Biblis-B, Brunsbüttel, Grafenrheinfeld, KNK II, Mülheim-Kärlich, Neckarwestheim-1, Philippsburg-1 & -2, and Unterweser.
• Eleven reactors are in the ease-off stage: Greifswald 1–5, Gundremmingen-A, Isar-1, MZFR, Obrigheim, Rheinsberg, and Stade.

1. Overview of Decommissioning Progress for Reactors in Germany Closed, 2015–2023

Sources: Various, compiled by WNISR, 2025

Note: a - Last production

With the demolition of the cooling towers in early 2023, decommissioning is advancing at the Biblis plant.1089 Both reactors are defueled, and with the start of RPV dismantling, Biblis-B has moved to the hot-zone stage.1090

The 1410-MW reactor at Brokdorf was closed in December 2021. Although the first decommissioning license was granted only in October 2024,1091 preparatory work had been ongoing, including primary circuit decontamination which was completed in 2023.1092 In August 2024, operator PreussenElektra (a subsidiary of utility E.ON) applied for the second stage of decommissioning licensing to begin hot-zone works. Defueling of the site is scheduled for completion in 2025.1093

The decommissioning license for the Emsland reactor, one of the last three reactors to close in Germany, was granted in September 2024. Since the reactor’s closure in April 2023, preparatory work had already been ongoing, with a full system decontamination of the reactor pressure vessel and connected systems being completed in early 2024. Furthermore, spent fuel has been removed from the reactor core and placed in wet storage.1094 The Lingen reactor, located close by, was removed from 27 years of LTE in 2015, the license for the second stage of decommissioning was granted in 2021, and currently, preparations for RPV dismantling are underway.1095

The Grafenrheinfeld reactor was closed in 2015, and hot-zone dismantling is underway. Recently, all water was drained from the reactor internals and the former fuel-storage pond.1096 Primary circuit decontamination was completed in 2016.1097

The last fuel element was removed from Grohnde’s reactor core and placed into the spent-fuel pool in 2022 where it was to remain for three to four years until the fuel can be transferred into dry storage and the reactor classified as defueled. Full system decontamination at Grohnde, which closed in 2021, was completed in 2023.1098

The reactors B and C of the Gundremmingen plant are planned to be dismantled in parallel, despite them closing several years apart, i.e., in 2017 and 2021, respectively. Gundremmingen-B has been defueled, while this step is expected to be completed for Gundremmingen-C sometime in the mid-2020s.1099 The completion target date for the project is 2040. At both reactors, full system decontamination is yet to be carried out, and it is scheduled for 2027 and 2029, respectively.1100 Regardless, the dismantling of other components is underway, including generators, turbines, and other parts of the steam cycle, as well as parts of the fuel storage pond of Gundremmingen-B.1101 Furthermore, in March 2025, operator RWE (Rheinisch-Westfälisches Elektrizitätswerk) announced plans to demolish both cooling towers in the autumn of the same year,1102 but as of June, a date had not been published.1103

The power plant Isar (also referred to as Ohu) houses two reactors. The 878-MW BWR Isar-1 was closed in 2011 and has moved to the ease-off stage since WNISR2024. The 1410-MW PWR Isar-2 was closed in April 2023 and was part of Bavarian Prime Minister Markus Söder’s (CSU) nuclear restart ambitions in the run-up to the federal election.1104 Isar-2 received its decommissioning license in March 2024,1105 and the full system decontamination was completed that same year.1106

Krümmel was closed in 2015, but the decommissioning license was granted only in 2024, which allowed for reactor-internals dismantling preparations to begin.1107 Several preparatory tasks (e.g., defueling) had been completed before the approval of the license, including full system decontamination in 2016.1108

The Neckarwestheim site houses two PWRs. Neckarwestheim-1 was closed in 2011. The reactor is currently undergoing hot-zone decommissioning work, with the RPV having been removed and current demolition work focusing on the reactor building itself. At Neckarwestheim-2—one of the reactors to close in April 2023—decommissioning began right after closure with the approval of the decommissioning license. The primary circuit was decontaminated in 2023,1109 and as of 2025 further work is ongoing, such as cutting the reactor cooling system pipes.1110

Since WNISR2024, Obrigheim, which closed in 2005, has moved to the ease-off stage, with work now focusing on decontaminating the buildings. This task is scheduled for completion before 2030, followed by the demolition of the remaining infrastructure to a greenfield site.1111

Two reactors—one BWR and one PWR—are being decommissioned at the Philippsburg plant. Philippsburg-1 was closed in 2011 and remains in the hot-zone stage, with RPV dismantling underway. Philippsburg-2 was closed in 2019, with the primary loop fully decontaminated one year later. Here, too, the RPV is currently being dismantled. The end of regulatory oversight is envisioned for early- and mid-2030s, respectively.1112

Taiwan

On 17 May 2025, the operations of Taiwan’s last reactor, Maanshan-2, ceased,1113 enforcing the “nuclear-free homeland” (or “non-nuclear homeland”) policy.1114 This makes Taiwan the first Asian country to have closed all of its commercial nuclear power plants, and the fifth country worldwide, after Germany, Italy, Kazakhstan, and Lithuania. The Nuclear-Free Homeland policy first appeared as a principle in the Basic Environment Act of 2002, but the legally binding deadline of 2025 was only enshrined in 2017 (before being reversed by referendum in 2018).1115 Refer to the Taiwan Focus for more details on Taiwanese nuclear policies and developments.

With the closure of the Maanshan-2 reactor, the Taiwanese fleet of six closed reactors amounts to 5.05 GW. The Chinshan (or Jinshan) site hosts two 604-MW BWRs that generated their last power in 2014 and 2017, respectively. In accordance with Taiwanese regulations, the decommissioning permits were submitted three years before plant closure and were approved by the Taiwanese Nuclear Safety Commission (NSC) in July 2019.1116

Decommissioning at Chinshan is planned in four stages. In Phase 1, warm-up stage activities are to be conducted. Phase 2 consists of hot-zone tasks, such as reactor internals dismantling, and is planned to begin in 2026 and go on for 12 years. Phases 3 and 4, scheduled from 2038 onwards, include final decontamination and demolition work for the site to be released in 2043. Given the unavailability of a final waste storage facility, spent fuel shall be stored at onsite interim storage facilities planned for each plant site. The construction of such a facility was completed in 2013 at Chinshan.1117 However, for eleven years, a legal dispute between the operator Taiwan Power Company (Taipower) and the government of New Taipei City, which refused to approve operator’s water and soil conservation plans, had hindered its commissioning. In April 2024, both parties came to an agreement according to which Taipower would improve the safety of the facility’s outer walls and its drainage system.1118 The final certificate was issued in October 2024, and after initial hot testing until December 2024, Taipower began transferring spent fuel from the ponds into dry storage.1119 The site is considered in the warm-up stage. Per the latest reported plans, defueling is to be completed in early 2026.1120

At the Kuosheng power plant, which houses two 985-MW BWRs that were closed in 2021 and 2023, respectively, spent-fuel pools have almost reached their full capacity, delaying continued defueling of the reactor core until a dry storage facility becomes available onsite.1121 As with the facility at Chinshan, the government of New Taipei City had concerns regarding the site’s water and soil conservation plans. The legal conflict was mediated before the Taipei High Administrative Court in August 2024,1122 and so finally, groundbreaking for the construction of such a facility could begin in December 2024.1123 Taipower anticipates finalizing construction by 2026 but does not expect to obtain an operating license allowing it to commission the facility before 2027.1124 With reactor dismantling to begin in 2029, overall decommissioning activities are scheduled to conclude with full site remediation in 2048.1125 However, as of mid-2025, the final decommissioning permit had not yet been granted,1126 and the delay regarding the onsite dry waste storage facility could lead to further delays in the whole project.

The Maanshan site houses two PWRs with capacities of 936 and 938 MW, respectively. Maanshan-1, the smaller unit, closed in July 2024 and Maanshan-2 followed in May 2025. A joint decommissioning permit was submitted by Taipower in 2021. The environmental impact assessment, to be approved by the Ministry of Environment, is still under review.1127 Preparations for dismantling, such as radiation characterization surveys are currently being conducted.1128 WNISR thus counts the reactors as being in the warm-up stage. At Maanshan, the construction of an onsite dry storage facility is planned with an expected startup “five to six years” from the closure of Maanshan-1.1129

Conclusion on Reactor Decommissioning

Assuming a 40-year lifetime—and given that the current average world fleet age is 32.5 years—a further 93 reactors will have been closed by 2030 (reactors connected to the grid between 1985 and 1990), and an additional 158 will be closed by 2065 (reactors connected 1991–2025). This does not account for the 157 reactors that have already been operating for more than 40 years, an additional 33 reactors in Long-Term Outage (LTO), and the 63 reactors under construction as of mid-2025. It should be noted, however, that the general trend now is to extend operational lifetimes beyond 40 years, often with the aim of 60 years and in some cases 80 years. As shown in previous WNISR editions, the financial and technical challenges of reactor decommissioning are often underestimated. With more and more reactors reaching the end of their lifetimes, this underestimation will likely bring costly consequences.

Between 1 July 2024 and 1 July 2025, five reactors were closed in the world: Doel-1 (445 MW) in Belgium, Pickering-1 and -4 (515 MW each) in Canada, and Maanshan-1 and -2 (938 MW each) in Taiwan.

Worldwide, as of mid-2025, 218 nuclear power reactors have been closed, corresponding to over 109 GW of permanently retired capacity. Only 23 of these have been fully decommissioned, although some are still awaiting release from regulatory control, and only nine have been returned to greenfield conditions, meaning the sites are available for unrestricted use. An additional 151 reactors are in some state of decommissioning—of which three units in the U.S. might restart—while 44 reactors are in a long-term enclosure (LTE) state.

In the European Union, the 132 closed reactors represent 61 percent of the world’s total and decommissioning efforts are advancing sporadically. Germany, having closed its last three operating nuclear power plants in April 2023, faces the unprecedented parallel decommissioning of 31 reactors, with one additional reactor remaining in LTE. The U.K. is still implementing its new “lead-and-learn strategy” to its legacy fleet and is facing potential financial shortfalls for the decommissioning of its AGR fleet. In France, with the exception of GCRs, decommissioning is advancing to some extent, although completion dates for ongoing projects are gradually pushed back year on year, and cost projections rise continuously with their total amount remaining uncertain.

The only countries to have fully decommissioned any commercial power reactors remain the U.S. (17), Germany (4), Japan (1), and Spain (1). The latest addition to the list is the 141-MW José Cabrera reactor, also known as Zorita, located in the Spanish province of Guadalajara. This reactor was connected to the grid in 1968 and closed in 2006. Technical decommissioning was completed in September 2023, and the site is awaiting its release from regulatory oversight following greenfield remediation.

Most of these decommissioned reactors are of low power capacity, many of them first-generation designs, with an average capacity of 350 MW. Decommissioning operations of completed projects lasted over 20 years on average, sometimes years longer than operation.

The five reactors closed over the year since mid-2024 have entered the warm-up stage. In the U.K., 19 reactors are currently in the warm-up stage, trumped only by 25 reactors in Japan; in total, 96 reactors are in this stage. Two Russian reactors, Novoronezh-1 and  -2, moved from LTE to the warm-up stage, while the Italian reactor Trino was erroneously classified as being in the hot-zone stage in last year’s report.

Five reactors moved from the warm-up to hot-zone stage, namely Biblis-B in Germany, Hamaoka-2 in Japan, Mühleberg in Switzerland, and both reactors at the Trawsfynydd plant in the U.K. Thus, 33 reactors, located in the U.K. (11), Germany (10), the U.S. (6), France (3), Italy (1), Switzerland (1), and Japan (1), are currently in the hot-zone stage.

Several reactors have advanced to the ease-off stage. Most notably, Swedish plants Barsebäck and Oskarshamn completed hot-zone works, and a fifth reactor, Ågesta, also moved on to the ease-off stage. In Germany, dismantling work advanced at Isar-1 and Obrigheim. The U.S. reactor GE Vallecitos had been mistakenly categorized as being in LTE in previous WNISR editions and is now counted in the ease-off stage. This places a total of 24 reactors in the ease-off stage. Germany leads in this category with a total of eleven reactors, followed by Sweden (5), the U.S. (5), Slovakia (2), and Belgium (1).

As no additional reactor has been placed into LTE since WNISR2024, the number of reactors in LTE has dropped from 45 in 2024 to 42 in 2025 (+2 in warm-up, +1 reclassification).

Potential Newcomer Countries

This chapter provides an overview of countries that have started building their first nuclear power plant or have engaged in a process aimed at the introduction of nuclear power in their respective countries. There are only three countries that do not yet operate nuclear power reactors but have invited a foreign technology provider to build them: Bangladesh, Egypt, and Türkiye. In all three cases, it is the Russian Rosatom implementing the projects.

There are many other countries that have made declarations about their ideas, intentions, strategies, plans, or even policy decisions. Most of them are too far from implementation to be covered in an annual report covering the industry. The last section provides seven examples with a wide range of progress towards national nuclear power programs.

Potential Newcomer Countries with Active Reactor Construction

Bangladesh

Russia’s Rosatom is building two VVER-1200 nuclear reactors at Rooppur in Bangladesh since November 2017 and July 2018, respectively.1130 In July 2018, Rosatom announced that commercial operations were to commence in 2023 and 2024 respectively.1131 Since then, there have been delays due to a variety of reasons and it is not clear when the reactors will be commissioned.

In April 2024, one of Bangladesh’s largest newspapers reported that “the deadline for the project’s completion has been extended to 2027.”1132 Completion of the project, including the second unit being connected to the grid, is still expected to be 2027 according to reports after the transmission line was commissioned.1133

The project cost is estimated at US$12.65 billion, but 90 percent of this figure (US$11.38 billion) comes from a loan from the Russian government; the Government of Bangladesh is only providing the remaining 10 percent. Of this, Russia had reportedly released US$7.33 billion by 14 August 2024.1134

The agreement between the Bangladesh Atomic Energy Commission and the JSC Atomstroyexport of the Russian Federation, signed on 25 December 2015, reportedly “requires semiannual payments of the loan, due on 15 March and 15 September. Each year, [US]$379.33 million is to be paid towards the principal amount, with [US]$189.66 million per instalment.”1135 But as reported in WNISR2024, due to economic challenges, the government was seeking a two-year extension on repayment of the loan from Russia for Rooppur.1136 According to media reports, “Although the principal repayment was initially scheduled to begin in March 2027, Russia has agreed to defer it by 18 months at Bangladesh’s request. However, Russia continues to press for regular interest payments as per the agreement.”1137

The other problem with repaying this loan results from Russia being forced out of the SWIFT system following its attack on Ukraine. As a consequence, Bangladesh could not send payments to Russia through this system. According to news reports, in 2023 Bangladesh agreed to repay in Chinese yuan through a bank in China.1138 However, that does not seem to have happened, and in 2024, Russian authorities suggested having Bangladesh open a branch of a Russian bank in Dhaka, but some commentators pointed out that this would be problematic for Bangladesh due to Western sanctions.1139

Finally, in December 2024, the two countries reportedly came to an agreement that would see the Russian government open a foreign currency account in Bangladesh’s state-owned Sonali Bank in Dhaka.1140 It is not clear if this arrangement actually works as in March 2025, there were media reports about blocked Sonali Bank payment attempts,1141 and in June 2025, it was reported that Bangladesh “owes around $700 million in interest arrears, which remain unpaid due to sanctions-related complications.”1142

The high cost of Rooppur has also led to accusations about corruption and graft.1143 In December 2024, Bangladesh’s Anti Corruption Commission (ACC) reportedly “alleged that there were financial irregularities worth about $5 billion” in the Rooppur project alone—besides a whole series of further corruption accusations—involving former Prime Minister Sheikh Hasina and members of her family, including Tulip Siddiq, her niece and now-former British Treasury Minister; prompting the ACC to launch an investigation.1144

In December 2024, Rosatom announced that the first unit had been fully built1145 and a Bangladeshi official confided in The Daily Sun, he was hopeful that it “would go online by March [2025].”1146 But at the end of February 2025, the head of Rosatom, Alexey Likhachev, could only announce that “the test run will commence soon.”1147 And a couple of weeks later, Likhachev vowed that “generating the first kilowatt-hours from the new nuclear power plants in Turkey and Bangladesh… are priorities for the current year.”1148

Building the transmission lines to ship the electricity from the plant had been seen as a problem for long. The likelihood of delays to transmission from Rooppur was identified back in 2020, and an official from the Power Grid Company of Bangladesh warned even then against the plant not being “able to start operation even if the nuclear power project is completed in time” if “the power transmission infrastructure and facilities are not ready.”1149 However, in early June 2025, the Power Grid Company of Bangladesh reportedly commissioned the transmission line, technically allowing electricity from Rooppur’s first unit to be exported.1150

Thus, the main reason for delays in commissioning the project now falls back to the construction of the reactor itself. There are still a number of technical challenges. According to a professor in the Department of Nuclear Engineering at Dhaka University, some of the challenges are “incomplete safety tests and compliance procedures required for power start-up, uncertainty regarding actual project costs impacting the finalization of the power purchase agreement, lack of necessary gridlines [Editor’s note: statement precedes powerline completion], the preparedness of certified reactor operators, and the absence of an established emergency preparedness and response centre.”1151

Adding to this are complaints by workers and engineers about working conditions, reportedly raising concerns about safety and security. Between 28 April and 6 May 2025, a number of employees of the Nuclear Power Plant Company Bangladesh Limited (NPCBL) went on strike. According to media reports, this resulted in 18 engineers being dismissed from their positions.1152

Bangladesh’s economic challenges, including these large payments that it will have to pay, are evidently taking a toll on investing in renewable energy. The capacity of renewables increased by only 100 MW during 2024; the total capacity at the end of 2024 was 1.1 GW.1153 The increase was entirely due to solar energy expanding by 13 percent, from 746 MW in 2023 to 846 MW.

Egypt

Egypt’s first nuclear power plant at El Dabaa involves four VVER-1200 reactors that have been under construction since July 2022 (Unit 1), November 2022 (Unit 2), May 2023 (Unit 3), and January 2024 (Unit 4).1154 The project is being financed largely through a US$25-billion loan from Russia—covering about 85 percent of the estimated costs—and Egypt is to repay this over 22 years with a 3-percent interest rate, starting in October 2029.1155

As discussed in previous WNISR editions, the long history of Egypt’s nuclear ambitions stems back to the mid-1950, and the El Dabaa site was already selected in the early 1980s.1156 The current US$30 billion-project is implemented by Russia’s state-owned company Rosatom and its subsidiaries Atomstroyexport, TVEL, NFCL, and Rusatom Service, as provided by contracts signed in 2015 and put into effect in late 2017.1157

The project has been delayed (see past WNISR editions for further detail). In December 2017, when notices to proceed were signed, the first unit was to be commissioned in 2026.1158 Rosatom maintained in March 2021 that the “launch” of the first unit was “scheduled for 2026” and all four units would be “completed by 2029”.1159 By May 2022, it was anticipated that the first two units would start operating in 2028 and 2029 respectively, and full operation of the plant would occur in 2030.1160

Per latest available information, these targets remain valid despite Egypt’s apparent eagerness to speed up the process. A Rosatom Newsletter paraphrased Madbouly, as “reiterat[ing] Egypt’s commitment to providing all possible assistance to accelerate the project delivery in cooperation with Rosatom” in August 2024.1161 In November 2024, during a meeting between Egypt’s Prime Minister Mostafa Madbouly and Rosatom CEO Alexey Likhachev, Madbouly reportedly “underscored Egypt’s commitment to completing the first phase of the project as scheduled,” while Likhachev “praised Egypt’s efforts to expedite the project.”1162 In February 2025, Likachev told TASS that Rosatom still expects to complete all four units by 2030.1163

Electricity in Egypt is generated mostly from natural gas (81.3 percent in 2024), followed by oil (7.5 percent), hydro power (6.1 percent), and wind and solar (5.1 percent).1164 Egypt’s renewable energy capacity has continued to grow, from 3.7 GW in 2015 to 6.7 GW in 2023 and 7.8 GW in 2024.1165 Of these 7.8 GW, wind energy constitutes 2.2 GW and solar energy 2.6 GW, which has grown dramatically from 36 MW in 2015. Egypt’s official target for renewables is to constitute 42 percent of its electricity generation mix by 2030, but officials have stressed that it would be contingent on international support.1166 The role nuclear is to play in this future strategy has not been disclosed yet, but according to some press reports, Egypt is considering to build more reactors, for which it is already setting land aside.1167

Türkiye

Türkiye has no operating reactor yet and one nuclear plant with four units under construction.

Unit 1 of Akkuyu nuclear power plant is currently in the final stages of construction and is scheduled to begin trial operation in late 2025, according to Turkish President Recep Tayyip Erdoğan.1168 As previously reported, the project has encountered delays at various stages (see earlier WNISR editions).

Originally, Unit 1 was planned to commence operations in 2023, with construction having started on 3 April 2018.1169 The other three units have been under construction since April 2020, March 2021, and July 2022 respectively.1170 Despite persistent delays and financing challenges, the authorities, including Türkiye’s Minister of Energy, Alparslan Bayraktar, continue to insist that all four units of the plant will be completed by 2028.1171 Nevertheless, skepticism remains.

As reported in WNISR2024, the schedule for the commissioning of Unit 1 had previously been revised several times between October 2023 and June 2024, then targeting late 2025 as startup date. In response to a parliamentary question submitted by Zonguldak MP Deniz Yavuzyılmaz in February 2025, Bayraktar maintained that Unit 1 would be commissioned in 2025, with subsequent units being commissioned at one-year intervals. This timeline suggests that the final unit is expected to be operational by 2028.1172 Having said that, Anton Dedusenko, chairman of the Board of Directors of Akkuyu Nuclear JSC gave 2026 as a launch date in an interview in early July 2025.1173

While delays to the original schedule appeared shortly after construction start—long before the impact of the full-scale Russian invasion of Ukraine—Alpaslan Bayraktar attributed the latest delay in the construction of the first reactor at Akkuyu primarily to Siemens Energy’s failure to deliver contracted equipment. He further noted that the company had not provided a satisfactory explanation for the disruption.1174 Following the minister’s remarks, in a written response to German media Deutsche Welle, Siemens Energy cited export and customs restrictions arising from sanctions against Russia as the reason for deliveries to Akkuyu being halted for nearly one year.1175 Meanwhile, in response to MP Sezgin Tanrıkulu’s written parliamentary question regarding the delivery, Energy Minister Alparslan Bayraktar stated that Siemens Energy had not delivered some of the parts in question and that they had to be sourced from a Chinese company instead.1176

Akkuyu Nükleer A.Ş., the builder-owner company in which Rosatom is the majority shareholder, offered more specific insight, claiming that Siemens Energy had been contracted to supply switchyard equipment with a voltage of 400 kilowatts for power transmission. While approximately 40 percent of the deliveries had been completed, the remaining 60 percent were withheld in Germany. Consequently, the installed equipment had to be dismantled and was replaced with new parts from China.1177 Energy Minister Bayraktar confirmed that shipment of the Chinese equipment began in August 2024 and was then expected to be delivered in full by December 2024.1178 Later, in response to a written parliamentary question on 11 June 2025, he stated that not all of the equipment had been delivered, but it would be delivered and assembled in stages.1179 In early 2025, Rosatom CEO Alexey Likhachev announced plans to initiate legal proceedings against Siemens Energy for failing to deliver prepaid equipment.1180

Additional factors have also contributed to construction delays, including liquidity constraints that reportedly led to worker strikes over delayed wage payments1181 and escalated into the alleged dismissal of three thousand subcontractors as of February 2025.1182

The cash problem indeed deepened in 2024, as U.S. banks blocked some of the money transfers—worth at least US$2 billion—to Türkiye1183 after the Justice Department reportedly discovered that the Akkuyu project was used to circumvent sanctions and funnel funds from the U.S. to “bankroll Russian state initiatives”, per the Wall Street Journal.1184 Reports suggest that the Russian National Wealth Fund may intervene to solve Akkuyu’s financing problems.1185 However, a parliamentary question about the Fund’s involvement submitted by Istanbul MP Sezgin Tanrıkulu on 7 March 2025 to Energy Minister Alpaslan Bayraktar was not answered. Tanrıkulu also asked about the veracity of media reports that there had been layoffs at the Akkuyu site due to the U.S. freezing the US$2 billion loan package.1186 In his response, Energy Minister Bayraktar said that personnel planning at the construction site is the responsibility of the project company, but he did not provide any particular information on layoffs.1187

Mid-2025, a memo from TSM Energy, one of the main contractors at the power plant, instructing workers to use their existing leave and then take unpaid leave, was leaked on social media.1188 The news of layoffs, which first appeared in the media at the beginning of the year, then received more detailed coverage. It was reported that the number of employees at the plant dropped from 35,000 to 12,000, with 10,000 of the 14,000 Russian employees returning home because they had not been paid.1189 It is said that the financial resources that cannot be transferred to the project amount to US$7 billion.1190

10,000 of the 14,000 Russian employees returning home because they had not been paid

Rosatom’s intention to sell a 49 percent stake in the Akkuyu project has raised questions about financial difficulties. In the early stages of the project, selling a 49 percent stake to a consortium of three parties (Cengiz Holding, Kolin İnşaat Turizm Sanayi ve Ticaret, and Kalyon İnşaat) was on the agenda. However, the sale did not go through at that time. The same companies are now being reconsidered.1191

In response to the sanctions, Russia reportedly expressed the idea of paying Türkiye with natural gas.1192

Under the terms of the intergovernmental agreement, the delay at the Akkuyu power plant opens the door to price adjustment according to the conditions of the Power Purchase Agreement (PPA), unless force majeure is invoked.

According to Article 6, Paragraph 2 of the bilateral agreement:

The Project Company, with the full support of the Russian Party, shall put Unit 1 into commercial operation within seven years from the date of issuance of all documents, permits, licences, consents and approvals necessary for the start of construction of the NPP.1193

As construction officially began on 3 April 2018, more than seven years have passed since.1194 The agreement also stipulates that the remaining units should be commissioned at one-year intervals, following the commissioning of Unit 1.

Construction

Despite financial difficulties, Rosatom continues to report some progress on construction activities at the Akkuyu site. Rosatom announced in early 2025 that the first backup diesel generator was commissioned at Unit 1 and startup work of the pumping station was underway.1195 Sergei Butckikh, the General Manager of Akkuyu Nuclear JSC, announced that the first shipment of nuclear fuel for Unit 2 has been delivered.1196

According to the Minister of Energy, more than 350 Turkish engineers who received training in Russia have begun working at the Akkuyu plant.1197

As the project nears completion, strategic and geopolitical considerations have gained prominence. Russia’s Minister of Emergency Situations, Lieutenant General Aleksandr Kurenkov proposed conducting a joint exercise in 2026.1198 “During the exercise, Russian and Turkish experts will provide training on firefighting and managing emergencies involving the release of radioactive materials,” Kurenkov stated. The stated aim of the exercise is to enhance the safety of power plant personnel and nearby communities.1199 However, concerns persist over the broader geopolitical implications of the project. Some argue that the Akkuyu nuclear power plant could become a more contentious issue between NATO and Türkiye than the previous S-400 missile defense system controversy.1200

The European Union’s Türkiye 2024 Report also drew attention to Türkiye’s energy dependence on Russia, emphasizing the need for legislative alignment regarding nuclear safety and stress tests at the Akkuyu nuclear power plant.1201

Corruption Charges on the Akkuyu Nuclear JSC Board of Directors

There have been several changes in the Akkuyu Nuclear JSC Board of Directors approximately five months after the departure of Anastasia Zoteeva and Gennady Sakharov (see WNISR2024). The new board consists of four people: Anton Dedusenko, Henri Edouard Proglio—former President of French national utility EDF and the only non-Russian member—Ekaterina Lyakhova, and Ruslan Baychurin.1202 Gennady Sakharov, then-Director of Capital Investments, State Construction Supervision and State Expertise of Rosatom, was arrested on 28 March 2024 on bribery charges and sentenced by a Russian court in May 2025.1203 The only official comment on the matter in Türkiye came in October 2024 from the Energy Minister, Alpaslan Bayraktar who, in a written response to a parliamentary question, stated that the accusations against Sakharov were unrelated to the Akkuyu project or his previous role as a board member.1204

Other Nuclear Newbuild Plans

At COP29 in Azerbaijan, Türkiye signed the—unrealistic (see past WNISR editions)—international pledge to triple operating nuclear energy capacity by 2050.1205 In addition to Akkuyu, Türkiye officially plans to construct two large-scale nuclear power plants, drawing the attention of several international nuclear energy companies. Sinop, located in northern Türkiye has long been considered a potential site (see previous WNISR editions), while Kırklareli in the Thrace region has recently been mentioned more frequently by Turkish officials. Russia, South Korea, and China are engaged in negotiations with the Turkish government, and Canada recently joined these countries, expressing interest in supporting Türkiye’s nuclear expansion.1206 In response to a citizen’s inquiry filed under the Right to Information Act concerning the proposed nuclear power plant in Thrace, authorities confirmed that consultations were ongoing with China. For Sinop, Rosatom CEO Alexey Likhachev claimed to be one step ahead as President Erdoğan had reportedly extended an offer for Rosatom to build a second nuclear power plant there.1207 As of yet, however, none of the interested countries have officially committed to either of the proposed locations.

Nuclear Waste

Türkiye currently lacks a designated final repository for nuclear waste, and limited information has been disclosed to the public on this issue. Responses obtained through Freedom of Information requests indicate that spent nuclear fuel will be stored onsite at the plant for the duration of its operational life.1208 In its response the regulator (NDK) indicates that under current plans, “the Project Company (operator) is responsible for the safe management of spent nuclear fuel and radioactive waste arising from the operation of a nuclear power plant and for the decommissioning of the facility.” According to the National Radioactive Waste Plan (URAYP), a geological disposal facility is projected to be commissioned by 2085.1209

Meanwhile, a Near Surface Disposal Facility for low and intermediate level waste, is scheduled to become operational by 2035. Additionally, the long-term expansion of the solid and solidified waste storage areas at Akkuyu is under consideration to accommodate waste generated over the plant’s anticipated 60-year lifespan.1210

Energy Outlook

As of mid-2025, Türkiye’s total electricity generating capacity reached over 119 GW.1211 Hydroelectric power plants lead with 32 GW, followed by natural gas power plants (25 GW), solar energy (23 GW), and wind energy (13 GW).1212 Power plants fueled by both domestic and imported coal together account for 21 GW of capacity. In the preceding 12 months, 5 GW of solar and 1.2 GW of wind capacity were added to the system while other sources remained largely unchanged, with only minor increases in capacity.1213 Solar and wind energy have been the leading players in Türkiye’s capacity growth in recent years, and this development has been reflected in electricity generation. Solar power’s contribution to electricity generation in Türkiye went from only 3.6 percent in 2020 to 7.5 percent in 2024; thus the solar share more than doubled in just four years. Wind energy’s share rose from 8.1 percent to 10.7 percent, and biomass also increased during the same period from 1.4 percent to 2.6 percent.1214

Peak electricity demand was recorded at 58.7 MWh on 23 July 2024 at 3:45 PM.1215 According to the Turkish Electricity Transmission Corporation (TEİAŞ)’s demand forecast report covering the period between 2025 and 2034, the peak electricity demand is expected to reach 81 GW by 2034 based on the reference scenario. The same scenario sees Türkiye’s electricity consumption going from 348 TWh in 2024 to 478 TWh in 2034.1216

Türkiye’s total greenhouse gas emissions as CO2 equivalent were 552 million tons at the end of 2023.1217 Türkiye’s Nationally Determined Contribution (NDC) target for 2030 is to keep total emissions below 695 million tons of CO2 equivalent.1218 Although this target implies an increase from the 2012 base year, Türkiye’s aim is to increase the capacity of renewable energy sources in order to limit emissions compared to a “business as usual” scenario. According to the National Energy Plan of Türkiye, the country’s energy goals for 2030 are to reach 33 GW of solar capacity, 18 GW of wind capacity, 35 GW of hydroelectric capacity, and 4.8 GW of nuclear capacity (corresponding to the Akkuyu plant).1219 The Energy Transition–Renewable Energy 2035 roadmap, which was announced at the end of 2024, set out a very ambitious target of 120 GW of combined installed wind and solar power capacity by 2035.1220

Potential Newcomer Nuclear Countries in Africa

African countries have some of the highest population growth rates in the world, and this is coupled with substantive economic growth and an increasing middle class that have resulted in dramatically higher electricity demand and associated acute power supply disruptions.1221 Interventions that show promise of alleviating the electricity shortfalls enjoy interest and support from the political leadership and the public at large.

As a result, initiatives to build nuclear power plants have been common in many African countries over the past decades. Importantly, outside South Africa and Egypt—who already operate or are in the course of building a nuclear plant—only four African countries having considered nuclear power (Algeria, Libya, Morocco, and Nigeria) have an operational capacity on the grid exceeding 10,000 MW. These are therefore the only countries in which building a typical 1,000 MW nuclear power generating unit would not exceed 10 percent of the total national power generating capacity, an order of magnitude considered a maximum, including for the IAEA. See Africa Focus in WNISR2024 for details and references.

WNISR2024 also detailed efforts over many years by the Russian state and its nuclear utility Rosatom to very actively seek nuclear cooperation agreements, typically with a statement of intent to eventually build a nuclear plant in any interested nation, regardless of that country’s financial capacity.

While, until recently, Russia had been largely unchallenged in pursuing aggressive nuclear marketing in Africa, there were already signs during the period covered by WNISR2024 that China was increasing its presence in Africa. A London-based China expert recently stated: “China is strategically positioning itself to exert significant influence over Africa’s nuclear energy sector in the coming decades.”1222 However, unlike Russia, Chinese efforts appear focused on aid in energy projects in general rather than specifically nuclear ones—its ability to supply solar panels and build solar farms fast and on a massive scale in particular ensured a significant competitive advantage.1223

The United States also responded to interest in nuclear technology, often in particular marketing Small Modular Reactors (SMRs). These efforts focused on economically stronger countries such as Ghana and Kenya, with which the U.S. concluded nuclear agreements in the past year.1224 It remains to be seen if the American initiatives will persist under the Trump administration, which has already instituted major cuts to aid and developmental initiatives in Africa.1225

Over the period covered by the current report, China has again substantially increased its activity in the African energy sector. It is now directly building large solar farms in countries such as Zambia1226 and Botswana.1227

When it comes to nuclear projects, China is now seemingly preferred to Russia in several countries where Russia was once considered the clear front-runner for partnering on a nuclear build. The three African countries that appear to have made some planning progress towards a nuclear power program have all had stronger engagement with China in the past year, as laid out in the sections that follow.

Opposition to nuclear developments in Africa has typically taken the form of local initiatives, but in June 2025, nine African organizations from Burkina Faso, Ghana, Kenya, Nigeria, South Africa, Tanzania, and Uganda met and cooperated with European organizations in the production of a report titled “The Alarming Rise of False Climate Solutions in Africa: The Nuclear Energy Misadventure”.1228 This joint analysis of the issues surrounding attempts to grow the nuclear sector in Africa could mark increased cooperation and effectiveness of anti-nuclear civil movements on the continent.

Ghana

Ghana has long explored a move to nuclear energy and has, as a result, already established associated governance structures, institutions, and processes. It has a Nuclear Regulatory Authority,1229 the Ghana Atomic Energy Commission with its Nuclear Power Institute,1230 and a company, Nuclear Power Ghana, set up in 2018 to develop and operate the country’s proposed first nuclear plant1231. It also runs a small research reactor.

It was announced in August 2022 that Ghana has officially decided to include nuclear energy in the planning for the future national electricity mix.1232 In September 2023, Nsuban, a coastal site, was designated as preferred location for the country’s proposed first nuclear plant, along with an inland backup site at Obotan.1233 In May 2024, it was reported that “16 countries and companies” had sought interest in contributing to a 1 GW nuclear plant, and that the list had been narrowed down to bidders from five countries: China, France, Russia, South Korea, and the U.S., with the Ghanaian authorities planning to make a decision on the successful bidder by December 2024, for the addition of 1 GW of nuclear power by 2034.1234

The nature and outcome of this bidding process is unclear. None of the major nuclear builders have communicated on the issue. At the IAEA’s General Conference in September 2024, Ghana stated it had “signed two Corporation Framework Agreements for a Small Modular Reactor (SMR) and a conventional large reactor. Successful negotiations with the two vendors will lead to the country potentially executing two projects concurrently.”1235

Some reports from March 2025 suggest that China had been identified as the preferred builder of the large plant, which would now have a capacity of 1.2 GW. At the same time a U.S.-Japanese SMR initiative led by NuScale Power and Regnum Technology Group would build twelve SMR modules of 77 MW each, totaling 924 MW.1236

Despite reports on such decisions, it is impossible to find official documentation, with the website of the designated operator of nuclear plants, Nuclear Power Ghana, silent on the issue.1237 The same reports however state that no contracts have been signed ratifying such agreements—in-line with Ghana’s statement to the IAEA General Conference—neither have transactional details such as costs and funding been disclosed. Adding to the uncertainty, there seem to be conflicting views regarding the preferred mode of nuclear power, with some government officials actively promoting a floating nuclear plant.1238

In February 2025, Ghana hosted an IAEA Site and External Events Design Review Service (SEED) mission to review the nuclear site selection process.1239 This illustrates that the process of identifying the potential plant location is still in progress, and that any construction start is still some way off.

Kenya

While talk about establishing a nuclear plant in Kenya has been going on for decades, and the country had instituted supporting steps such as setting up an official Nuclear Power Energy Agency (NuPEA), there had been few concrete moves to steer such a project towards realization. Towards the end of 2024 there were still plans to initiate by 2027 the construction of a 1 GW plant adjacent to the Indian Ocean at Uyombo in Kilifi County.1240 This in turn galvanized opposition from local communities, non-governmental organizations and parliamentary opposition.1241

In what would have been a far-reaching move, in January 2025 the Kenyan government announced its intention to dissolve NuPEA.1242 The motion did not specifically target NuPEA, as many other governmental organizations were equally affected. It also did not aim to shut down NuPEA’s functions, but rather to relocate them elsewhere in government.1243 This is however a clear indication of uncertainty and instability in the organization responsible for managing all potential nuclear programs, and, although NuPEA continues to function for now, it is difficult to see any nuclear developments proceeding in this state of flux. Notably, in its budget the Kenyan government only allocated 743.8 million Kenyan shillings (about US$5.7 million) to nuclear energy development for the 2025/2026 financial year.1244

Should a nuclear newbuild project be initiated, China here too has been trying to position itself as the favored construction partner, and an intergovernmental Memorandum of Understanding strengthening its partnership with Kenya was signed in March 2025.1245 The United States had also been actively pursuing nuclear development projects in Kenya in the second half of 2024, concluding a general nuclear cooperation agreement in August 2024.1246 It remains to be seen whether and to what extent such initiatives will be pursued under the new U.S. administration.

Unlike Ghana and Nigeria, Kenya has made some progress in the past year in initiating renewable energy projects to help alleviate power scarcity, although these developments are still not happening on the scale required to achieve medium term energy security.1247

Nigeria

For years, the Nigerian administration and various national institutions have strongly supported the idea of the implementation of a national nuclear power program. Nigeria had in prior years signed agreements on nuclear power development with South Korea, France, Russia, and India, and the Nigerian Nuclear Regulatory Authority (NNRA) signed cooperation agreements with nuclear regulators in the U.S., Pakistan, South Korea, and Russia.1248 Preferred locations for nuclear builds have reportedly been identified at Geregu, for which some preliminary feasibility studies have reportedly been carried out,1249 and Akwa Ibom, but no firm site decision and approvals appear to have been made.1250 A currently suggested timeline for building a local nuclear can be viewed in a presentation made in late 2024 by an official of the Nigerian Atomic Energy Commission, which proposes the construction of a nuclear plant to commence in 2028/2029 with commissioning in 2034.1251

As reported in WNISR2024, Chinese nuclear developers are now active in Nigeria, having opened a regional office in Lagos.1252 During a state visit to China in September 2024, Nigerian President Bola Tinubu signed a series of bilateral cooperation agreements, including cooperation in nuclear power developments.1253

In recent years projections for nuclear builds suggested a setup akin to the Akkuyu and El Dabaa projects, i.e. four units of 1200 MW each. Such a configuration is however not supported by the current Power Minister, who instead favored nuclear power generation using SMRs.1254

Despite the pronounced ambitions to soon develop nuclear plants and activities suggesting work on this is well under way, it is telling that the 2024 Nigeria Integrated Resource Plan, the country’s official roadmap for electricity generation, has not included any nuclear power in its projections leading all the way to 2045.1255

There is therefore no indication that Nigeria has in the last year moved any closer to initiating a nuclear power program.

Uganda

WNISR2024 reported on Uganda’s vague and unrealistic ambitions to develop 24 GW of nuclear power. This target was reaffirmed in 2025, which is being viewed as a major contributor to an aimed-for total generation capacity of about 52.5 GW by 2040,1256 which in turn corresponds to over 25 times the country’s mid-2024 grid generating capacity of 2 GW.1257 In line with the proposed first step of building a 1-GW unit of an eventually envisaged 8.4-GW nuclear plant at Buyende, the Ugandan government in May 2025 concluded a contract with Korean Hydro and Nuclear Power (KHNP), where this entity, with Dohwa Engineering and KEPCO E&C, would “undertake pre-feasibility or site evaluation studies” over 26 months, only after which a full feasibility study and resettlement plan would commence.1258 This schedule alone casts major doubt on the declared aim expressed by the responsible Ugandan Minister to have the first 1 GW unit at Buyende operational by 2031.1259

Case Studies Italy and Poland

Italy

Italy was one of the very early adopters of nuclear power and started building its first reactor, the 153-MW Gas Graphite Reactor at Latina, in 1958. It was commissioned in 1963 and was closed in November 1986. The last of four and largest unit to be commissioned in the country—the 860-MW BWR at Caorso—started building in 1970, was connected to the grid in 1978, and closed in October 1986. The 260-MW PWR Enrico Fermi (or Trino) was the last nuclear reactor to generate power in the country. It was first connected to the grid in 1964 and closed in March 1987.

The nation was deeply shocked about the Chornobyl disaster in April 1986 that also led to significant levels of contamination in Italy with parts of the harvest being destroyed. A national referendum in November 1987 led to the termination of Italy’s nuclear program.

When the Chornobyl disaster began, besides the four operating units, three additional reactors were under construction: the two 982-MW BWRs Montalto Di Castro-1 and -2 and the 35-MW prototype Heavy-Water Light Water-Cooled Cirene reactor. Their construction was abandoned following the 1987 referendum.

A second national referendum was held in June 2011, only three months after the Fukushima disaster started unfolding. Then-President Berlusconi had planned to reintroduce nuclear power and had passed legislation to allow for nuclear newbuild, but 94 percent of Italian voters rejected the law.1260

Decommissioning

Since the late 1990s, decommissioning at all four facilities has been underway and is in various stages with the final goal of brownfield release (i.e., completed decommissioning but wastes or other materials still onsite). The work is being conducted by Italian company Società Gestione Impianti Nucleari SpA (Sogin), which in 2024, 37 years after the last nuclear power generation, signed a collaboration agreement with the European Commission’s Joint Research Center “to implement and develop a common nuclear dismantling and radioactive waste management strategy.”1261

Decommissioning of the smallest reactor operated in Italy—the 150-MW BWR Garigliano—is underway, but when reactor dismantling is to be completed is unclear, as the Italian project website claims a target year of 2031,1262 while the English language site claims the year 2025.1263 At the Enrico Fermi plant, decommissioning work has been ongoing since 1999. There has been no official update on progress since WNISR2021, but according to a news report from February 2024, reactor internals dismantling has now begun, and the completion date has been postponed to 2036.1264

At Italy’s largest reactor, the 860-MW BWR Caorso, decommissioning began in 1999, but the dismantling of the reactor building’s external components only began in November 2024 with internal dismantling expected to begin in 2026 and to be completed by 2030.1265 The release of the site as brownfield is expected for 2031.1266 See section on Italy in Decommissioning Status Report and previous WNISR editions for more details on decommissioning.

Italy is also in the process of finding a final waste repository, with radioactive waste currently being stored at ten interim storage facilities spread across the country. A map of potential locations for the final repository was released in 2022.1267

New Program-Restart Initiatives

Italy’s National Energy and Climate Plan (NECP) as published in July 2024 contains the term “nuclear” 110 times, which is surprising for a country that abandoned nuclear power almost 40 years ago. The report sees,

great potential to develop new nuclear technologies for Italy to help revitalise nuclear energy, not only at international level but potentially at national level.

It states that,

the recovery in Italian territory of the production of energy from nuclear sources could play an important role [in the reduction of greenhouse gas emissions].

Furthermore, it informs that the government has set up a National Platform for Sustainable Nuclear (PNNS) and notes that

in 2050, in the “With nuclear” scenario, nuclear production covers around 11 % of the electricity demand. In addition to meeting increased demand, nuclear reduces the need for both CCS [Carbon Capture and Storage] natural gas generation from 11.5 to 4 TWh and bioenergy production with CCS from 12.5 to 6 TWh.1268

This scenario projects 8 GW of nuclear capacity operating by 2050, an ambitious goal, but it estimates feasibility at double that amount, i.e., 16 GW of nuclear newbuild. While Italian companies have contributed to the construction of nuclear power plants in various countries,1269 the last time they started building a nuclear reactor at home that eventually generated power was nearly 55 years ago. Thus, Italy is a true—potential—newcomer country.

In July 2024, Électricité de France (EDF) signed a non-binding MoU with four Italian companies, Edison, Ansaldo Energia, Ansaldo Nucleare, and Federacciai, “aimed at promoting cooperation in the use of nuclear energy to boost the competitiveness and decarbonisation of the Italian steel industry.” The agreement further prudently states that the signatories would “undertake to consider co-investment opportunities in new nuclear energy and, in particular, in the construction of Small Modular Reactors in Italy over the upcoming decade.”1270

A number of Italian startups are also involved in nuclear development. One example is Terra Innovatum, a company founded in 2021 that promotes a helium-cooled, graphite/beryllium-moderated design called SOLO, “from 1 MW to 1 GW”, and presents it as “The World’s First Micro Modular Nuclear Reactor commercially deployable anywhere globally within the next three years.”1271 In an April 2025 investor presentation, the company claimed to have “completed” the design in October 2024, but the design has not yet gone beyond pre-pre-licensing steps in the U.S., where Terra Innovatum in January 2025 filed a Regulatory Engagement Plan for a non-power reactor license application.1272 In the 13-page document, the company states its intention to apply for “a Construction Permit (CP) of the FOAK [First Of A Kind] Research Test Reactor (RTR) facility to be deployed on a site yet-to-be-determined in the US territory.”1273 The company has “2–10 employees”, according to its LinkedIn page.1274

Political, Legal, Industrial Steps and Comments

In December 2024, the 100% Renewables Network published a 12-page Nuclear Cost Report that concluded that “Nuclear would make electricity more expensive. This is much more expensive than renewable sources. And the ‘small’ ‘Small Modular Reactor’ reactors are even more expensive.”1275

On 28 February 2025, the Italian government nevertheless approved a draft delegation law on “sustainable nuclear energy”. The draft law demands “a clear break with respect to the nuclear plants of the past” and advocates for the “use of the best available technologies, including modular and advanced technologies.”1276

On 14 May 2025, the three companies Enel, Ansaldo Energia, and Leonardo established Nuclitalia, which “will be in charge of studying advanced technologies and analyzing market opportunities in the new nuclear sector” and “assessing the most innovative and mature designs of new sustainable nuclear power, with an initial focus on water-cooled Small Modular Reactors (SMRs).” Enel holds 51 percent of the capital, Ansaldo Energia 39 percent, and Leonardo 10 percent. All three companies are “under the common control of Italy’s Ministry of Economy and Finance.”1277

The establishment of the Eagles Consortium was announced on 16 June 2025. Italy’s Ansaldo Nucleare and ENEA got together with Romania’s RATEN and Belgium’s SCK CEN “to design and commercialize Lead-Cooled IV Generation Small Modular Reactor” and is committed to start up “its first demonstrator by 2035”. The reactor would have “around 350MWe power” and use uranium-plutonium fuel.1278

Specialized trade platform Montel so far sees that “Italy is taking small steps toward reviving its long-abandoned nuclear sector, with political support growing but lingering public resistance, high costs, and policy uncertainty still clouding its return.”1279

In May 2025, the daily il Fatto Quotidiano published a blog piece that stated:

The only way to manage diversity in the majority is a frenzied centralization of decisions in the hands of the prime minister. An example of this is the way Italy is dealing with the relaunch of nuclear power; journalists and complacent “experts” repeat Pichetto Fratin and Georgia Meloni’s statements without any comparison on the merit and consequences that the atomic fission processes will impose on the country’s energy, economic, social and environmental balances, exposing future generations to risk.1280

A June 2025 assessment by the Bank of Italy concludes:

The analysis shows that, given the structure of the market and the electricity bill, reintroducing nuclear power would not have a significant impact on price levels. (…) In terms of energy dependence, the reduction in hydrocarbon imports would be offset by increased imports of technology and fuel for nuclear production, which are currently concentrated in countries that are not particularly close to Italy in geopolitical terms. (…) Another important element that emerges from the analysis is the uncertainty surrounding the technologies chosen, most of which are not yet available for commercialization. This uncertainty calls for a cautious approach that also prepares and promotes alternative strategies.1281

Il Fatto Quotidiano commented on the Bank of Italy’s analysis by saying it “dismantles the pro-nuclear rhetoric”.1282

Electricity Mix

Electricity consumption has been relatively stable over the past 25 years, peaking in 2008 at 352 TWh and dropping to 315 TWh in 2024. Italy has constantly increased electricity imports since 2008 when net imports were just 3 TWh to 51 TWh in 2024.

In 2007, fossil fuels provided 84.4 percent of the domestic electricity generation and renewables 15.6 percent. In 2024, fossil fuels were down to 50.7 percent, of which 44 percent was natural gas, and renewables had more than tripled to 49.3 percent, of which hydro contributed 19.3 percent, solar 13.5 percent, wind 8.4 percent, and bioenergy 6 percent.1283

Poland

WNISR2024 featured a focus section dedicated to Polish nuclear ambitions. It covered various past implementation attempts and detailed how plans evolved. From the 2018-draft Energy Policy of Poland until 2040 to the latest Polish Nuclear Power Program released for public consultation in 2025, the extended target of deploying 6–9 GW of nuclear capacity has been the official policy of successive administrations in recent years.

As depicted in WNISR2024, ongoing and tracked activities can be divided into three areas:

• Construction of three large reactors at the Lubiatowo-Kopalino site in the Choczewo municipality in Pomerania, a project that was awarded to U.S.-based Westinghouse (as design provider of the AP-1000) following a tender which also included France’s Électricité de France (EDF) and South Korea’s Korea Hydro & Nuclear Power (KHNP).
• Construction of two APR-1400 reactors, a vague project, to be implemented by KHNP at a site yet to be determined.
• Various private sector driven initiatives based on Small Modular Reactors (SMRs), e.g. the BWRX-300 by ORLEN Synthos Green Energy (OSGE) with GE-Hitachi and projects by NuScale and Last Energy.

The Pomeranian Project (Lubiatowo-Kopalino)

Westinghouse was formally appointed in November 2022 as the technology vendor of three large reactors for the Pomeranian project at a price of around US$20 billion.1284 Westinghouse and the Polish state-owned project company, Polskie Elektrownie Jądrowe (PEJ),1285 signed a cooperation agreement in December 2022.1286 The companies further advanced the agreement in February 2023 by signing a contract covering front-end engineering, early procurement work, and program development.1287 In July 2023, the Ministry of Climate granted a “decision-in-principle” on the project allowing for further administrative applications to proceed.1288 In September 2023, PEJ and the newly formed consortium Westinghouse-Bechtel1289 signed an 18-month Engineering Services Contract (ESC) that was to provide, by the end of the period, the design documentation for the first nuclear plant in Poland.1290 This ESC expired in March 2025 but was replaced by an Engineering Development Agreement (EDA) between the three partners signed in late April 2025 to allow continuation of the project.1291

A few months prior to the EDA’s signing, new delays in the project were confirmed. While the start of construction work was scheduled for 2026 and electricity generation for 2033,1292 in December 2024 a three-year delay was announced.1293 According to the Secretary of State for Strategic Energy Infrastructure, Wojciech Wrochna, pouring of first concrete was then scheduled for 2028 and connection to the grid of the first unit in 2036, with the second and third unit to follow in 2037 and 2038.1294

Just after the ESC was signed, general elections in Poland resulted in a leadership change that brought a center-right government under Donald Tusk, who had already been Polish Prime Minister from 2007 to 2014, into office in December 2023.1295 While clearly in favor of nuclear power, prior to the election Tusk reportedly agreed that the Polish nuclear power construction plan was “not based on a robust economic analysis and lack[ed] a business plan.”1296 The new administration prompted an investigation into whether the target of first nuclear operations by 2033 was realistic. Incoming Climate and Environment Minister Paulina Hennig-Kloska said in February 2024 that the government doubted the ambitious target was achievable given the substantial delays and uncertainties.1297 Industry Minister Marzena Czarnecka stated in a radio interview in May 2024 that the earliest completion date of the Pomeranian project would be in 2039 or 2040.1298 After concerns were raised that the delay could lead to “an energy disaster”, Czarnecka clarified her statement saying that the target date of 2039 or 2040 referred to the whole plant’s completion and that the first reactor was to come online by 2035 with construction start in 2028.1299 As stated above, six months later, in December 2024, Wrochna reported another slightly different schedule with pouring of first concrete in 2028 and connection to the grid of the three units in the years 2036, 2037, and 2038.

In April 2024, PEJ’s then-authorized representative Jan Chadam said that the project’s final cost was still not confirmed and gave an indicative figure of PLN150 billion (~US$202438 billion),1300 but by the end of 2024 new estimates of “total investment costs” ranged around PLN192 billion (US$202448.2 billion),1301 more than double the estimates of November 20221302.

It should be noted that in recent projects Westinghouse refrained from taking on construction as the liable builder-company. In early June 2024, the Wall Street Journal reported: “Westinghouse said it learned from its U.S. experience during the 2010s and no longer takes on reactor construction…In Ukraine, Westinghouse said Energoatom will be responsible for construction of the new reactors…”1303 In Bulgaria, South Korean company Hyundai is the main contractor, and Westinghouse is providing technology and engineering. In Poland, Westinghouse has partnered up with U.S. construction giant Bechtel and the Polish state-owned PEJ.

Regarding the financing package, President Andrzej Duda signed a bill committing PLN60.2 billion (US$16 billion) from the state budget for 2025–2030 to cover 30 percent of the project’s estimated costs.1304 Additionally, the proposed state support mechanism includes state guarantees covering 100 percent of the debt taken by PEJ on the project and a two-way Contract for Difference (CfD) guaranteeing revenue stability over the entire 60-year projected lifetime of the plant. The bill was signed in March 2025.1305 The proposal led to the opening of an in-depth investigation by the European Commission in relation to:1306

• The appropriateness and proportionality of the aid package. Given there are three different aid measures (equity, guarantees, two-way CfD) that together limit the risk for the beneficiary, it is important to ensure that overall no more aid than what is strictly necessary is ultimately granted. In particular, the Commission will examine further (i) whether the 60-year duration of the CfD is justified taking into account the other two measures, and (ii) whether there could have been other companies interested in leading the project which might have resulted in a smaller aid amount.
• The impact of the aid package on competition in the electricity market and whether this is kept to the minimum. The Commission will investigate whether the design of the two-way CfD sufficiently incentivises the power plant to operate and participate efficiently in the electricity markets, within its technical capacity. This is important to minimise market distortions, facilitate the integration of renewables, and allow the electricity system to move towards decarbonisation. In particular, the power plant should plan its maintenance and refuelling in an optimal way and adapt its power production to market prices. In addition, the Commission cannot exclude that the aid will not be passed-on to electricity consumers through direct contracts.

In March 2025, the European Commission published an invitation for interested parties to submit comments to the state aid investigation.1307

Further financial aid would be provided by the U.S. International Development Finance Corporation (DFC), which signed a letter of intent in November 2024 to also provide PLN4 billion (US$1 billion) for the plant.1308 By June 2025, PEJ announced it had “obtained letters of support for more than 70% of the assumed amount of debt financing” from “11 export credit agencies from Europe, North America and Asia.”1309

PEJ experienced two major changes in the top management within two years. In February 2024, the whole management board was dismissed, and Leszek Juchniewicz was appointed as the new CEO.1310 Press coverage suggested that the move was due the board’s ties to the previous nationalist government. However, only a little more than a year after, in March 2025, Juchniewicz was also dismissed, with immediate effect, without grounds publicly provided.1311

A Second, Korean Project Remains Vague

In October 2022, a few days after Westinghouse was awarded the Lubiatowo-Kopalino project (see above), two preliminary agreements for the development of a nuclear power station based on Korean technology in Pątnów, central Łódzkie province were signed. Polska Grupa Energetyczna (PGE) and local electricity producer ZE PAK signed a letter of intent with KHNP to create a plan for the construction of the Pątnów plant, using Korea’s APR-1400 reactor technology, with a total capacity of 2.8 GW. Meanwhile Jacek Sasin, Poland’s then-State Assets Minister, and Chang-Yang Lee, South Korea’s then-Minister of Trade, Industry and Energy signed an agreement pledging cooperation in support of the project.1312

ZE PAK and PGE announced in March 2023 that they would establish a joint venture to “represent the Polish side at all stages of the [Pątnów] project”, which was planned with at least two APR-1400 reactors to be delivered by KHNP and scheduled to be on the grid by 2035.1313 In April 2023, the 50-50 joint venture, named PGE PAK Energia Jądrowa, was established.1314 In November 2023, the Ministry of Climate and Environment granted a “decision-in-principle” for two APR-1400 reactors. This approval marks a first step in the Polish licensing process for nuclear facilities.1315

In January 2024, KHNP president Joo-ho Whang visited Warsaw for the grand opening of a three-person office. He announced that an agreement for a feasibility study for the project was to be signed by the end of March 2024, and that the financing scheme and EIA would be concluded by 2025, allowing for an “ambitious but achievable” beginning of commercial operation of the first reactor by 2035.1316 However, in late May 2024—after the change in Polish government—PGE president Dariusz Marzec said that the company was “very far away” from making an investment decision, given that the ongoing pre-feasibility studies would require several years to conclude, casting uncertainty on the project’s specifics as defined so far.1317

While the Pomeranian project was followed by two engineering service agreements and concrete steps towards financing, this was not the case for the Pątnów project. With the change in government in late 2023, the momentum of the Pątnów project slowed down as well. In addition, Westinghouse disputed the export of APR-1000/APR-1400 technology by KHNP claiming license infringements, further escalating the conflict in the context of KHNP’s bid in the Czech tender (see Czech Republic).1318 However, in early 2025, South Korea and the United States signed a Memorandum of Understanding (MoU) notably “to cooperate in expanding civil nuclear power in third countries while strengthening their respective administration of export controls on civil nuclear technology.”1319 Only one week later, on 16 January 2025, Westinghouse announced a Global Settlement Agreement with KEPCO and KHNP ending the longstanding dispute over claims of unauthorized transfer of technology and technical information.1320 While the details of the agreement remain confidential, KHNP President Joo-ho Whang said that the agreement will serve as an opportunity to build closer cooperation between KHNP and Westinghouse.1321 But even with the resolution of the Westinghouse KHNP legal dispute, it remains to be seen if the resources to follow through with two large reactor projects in the country can be made available.

None of the steps taken so far are binding engagements. In January 2025, a preliminary agreement was signed between both Polish companies providing terms for the potential takeover by PGE of ZE PAK’s shares and equity interests in PGE PAK Energia Jądrow. Exclusive negotiation rights ran until 30 June 2025.1322 In its Group Strategy until 2035, released earlier that month, PGE announced “until 2035, expenditure only on preparation of projects for administrative and investment decisions,” clarifying that “subsequent decisions depend on the results of location research and market demand.”1323 The accompanying Press Release details “in nuclear energy area expenditures of approx. few hundred millions PLN solely for conducting necessary research and analyses in 3 potential locations – Bełchatów, Turów and possibly Konin.”1324

The Polish Nuclear Power Program released for public consultation by the Ministry of Industry in June 2025 expects site selection in 2027, where construction would start in 2032. It is now to host three units, which would be commissioned between 2041 and 2043. However, Poland seems to have hit reset on a possible cooperation with KHNP, as the provisional schedule for the second Polish nuclear plant project estimates that a “strategic partner (technology)” will be chosen in 2027. Curiously, a decision-in-principle is already expected in 2026.1325

Perspectives on SMR Deployment

When looking at plans for Small Modular Reactor (SMR) deployment in Poland—a newcomer country—one should keep in mind that the Polish nuclear regulatory environment was tailored to regulate only a few nuclear facilities. One requirement for the acceptance of the large reactor project at Lubiatowo-Kopalino, therefore, was a proven and already licensed design. The licensing of three AP-1000 reactors already required an expansion of staff at the Polish nuclear regulatory authority, PAA, and resulted in delays in routine tasks for existing facilities.1326 So, the licensing capacities for novel SMR designs involving a larger number of additional sites should be considered with caution. Nevertheless, Poland is eyeing the possibility of investing into the deployment of SMRs. Various cooperation agreements have been signed.

BWRX-300: In April 2023, the U.S. Export-Import Bank and the U.S. International Development Finance Corporation signed letters of interest to provide loans of up to US$3 billion and US$1 billion, respectively, to the ORLEN Synthos Green Energy (OSGE)-led project—jointly established by ORLEN and Synthos Green Energy—that as of April 2023 envisioned the construction of up to 20 GE Hitachi (GEH) BWRX-300 reactors in Poland, with the startup of the first one over-ambitiously scheduled for 2029.1327 In 2023, the BWRX-300 design passed an (optional) pre-licensing review in which the President of the Polish National Atomic Energy Agency issued a “general opinion on selected technical assumptions of the BWRX-300 reactor technology,” deeming them “compliant with the national nuclear safety requirements,” though pointing to one which “must be verified again before the start of the relevant administrative process related to obtaining licenses of the President of the PAA.”1328 In April 2025, the Canadian Nuclear Safety Commission (CNSC) issued a construction license for the BWRX-300 at the Darlington site in Ontario1329 (an informal prerequisite for a license in Poland). In November 2024, OSGE reached an agreement with Laurentis to cooperate on the preparation of the Preliminary Safety Analysis Report, which is now planned for mid-2026.1330

In September 2023, OSGE was also selected to receive funds from the U.S. Department of State for “Coal-to-SMR” feasibility studies under “Project Phoenix”.1331 On 7 December 2023, six projects—at Włocławek, Stawy Monowskie, Stalowa Wola, Ostrołęka, Nowa Huta, and Dąbrowa Górnicza—envisioning up to 24 individual reactors on the grid by 2030 (clearly unrealistic and overambitious), received approval-in-principle from Poland’s Ministry of Climate and Environment, only a few days before the new government came into power.1332

As of June 2024, three sites of the OSGE project, namely Stawy Monowskie, Włocławek, and Ostrołęka, totaling up to 4.6 GW or 15 reactors, had entered environmental proceedings with the General Directorate for Environmental Protection (GDEP) to establish the required scope of Environmental Impact Assessment (EIA) Reports, while applications had yet to be filed for the remaining three sites.1333 In February 2024, GDEP issued a decision on the required scope of environmental reporting on the Stawy Monowskie project and announced that it was starting to carry out environmental and siting studies necessary for the preparation of the EIA Report.1334 According to the company, it could take about two years for the EIA report to be completed.1335 In February 2025, GDEP issued similar decisions on the Ostrołęka and Włocławek project ideas, and OSGE can now begin environmental and siting research at those locations as well.1336

Last Energy: In June 2022, Polish state-owned company Enea S.A. and U.S. SMR developer Last Energy agreed to cooperate on the deployment of SMRs.1337 In March 2023, after contracting power purchase agreements (totaling over US$4.3 billion in electricity sales and US$1 billion in infrastructure investments) with industrial partners of Poland’s Katowice Special Economic Zone and in the U.K., Last Energy felt optimistic enough to announce the deployment of ten 20 MWe “Micro Modular Nuclear Power Plants” with the target of commissioning a first plant by 2025—later corrected on the website to 2027—to provide customers with electricity and heat.1338 The reactor design is not licensed anywhere yet, and a construction license application had not been filed as of mid-2025, which indicates that the delayed target date will not hold either.

NuScale: In September 2021, SMR developer NuScale signed an MoU with Polish mining company KGHM and engineering firm Piela Business Engineering (PBE).1339 In July 2023, the project was granted first approval by the Ministry of Climate and Environment, theoretically allowing the project to move to the next administrative steps, including siting decisions and building permits.1340 However, so far, the administrative procedures appear to be limited to so called “regulatory opinions”, a legal instrument in Poland to obtain a preliminary, non-binding, general assessment of a project design outside of the regulatory authority’s formal licensing process.1341

In March 2025, the Vice President of KGHM’s management board confirmed that SMRs were still considered an option “over the long term”; he did not mention NuScale, but he remarked that “the SMR project at KGHM has the status of a research and development initiative”, and provided the following reminder:1342

The future of all SMR projects in Poland will depend, among other factors, on the economic viability of such investments, progress in the commercialization of this technology abroad, and relevant legislation.

Conclusions on Nuclear Project Developments

Of the various Polish initiatives, two stand out and continue to move forward. The construction project of three AP-1000s at the Lubiatowo-Kopalino site and the OSGE BWRX-300 project. While the announced schedules for both projects (especially the BWRX-300) seem overly optimistic, both reactor designs meet the informal prerequisite of being licensed by an experienced regulator elsewhere. While the AP-1000 has been built in China and the U.S., the BWRX-300 has received a construction license in Canada. Both projects showed slow but consistent steps in the licensing procedure. This was not the case for the other projects. There was little to no movement from mid-2024 to mid-2025 in the NuScale and Last Energy projects, and nothing has been reported on an earlier Rolls-Royce SMR initiative (see Poland Focus: Perspectives for SMR Deployment in WNISR2024).

Non-Nuclear Energy Developments

Meanwhile, renewable energies are being rapidly deployed in Poland. In 2024, wind contributed 14.6 percent to total electricity generation, while solar contributed 9 percent—four times its share in 2021. The combined share of wind and solar doubled from 11.3 percent to 23.5 percent in just three years from 2021 to 2024. At the same time, the share of fossil fuels—still dominating the Polish power system—has been reduced from almost 83 percent to 70 percent, coal accounting still for 53.5 percent of total production, down from 71.3 percent in 2021.1343 In other words, while discussions about the nuclear program dominate the news, renewables have started to rapidly decarbonize the Polish power system.

Other Examples

There is an increasing number of countries that claim planning the introduction of nuclear power programs. Many of these claims are either lacking credibility or imply time frames that are at this point irrelevant for an annual report on the industry (see also Potential Nuclear Newcomer Countries in Africa). Early potential newcomers include Ecuador, Estonia, Indonesia, Jordan, Kazakhstan, Saudi Arabia, and Uzbekistan that are briefly analyzed hereunder.

Ecuador

According to the IAEA, “Ecuador is well on the road to including nuclear power in its energy mix.”1344 However, there does not appear to be any official energy plan including nuclear power so far. In October 2024, Ecuador’s President vaguely stated that there was a “dialogue with France” on nuclear power.1345

Later that month, it was reported that a Vice-Minister of Electricity had unveiled an “ambitious roadmap” to deploy a 300-MW reactor by 2029. A 1-GW reactor would follow in the longer term.1346 No official communication is available.

In May 2025, the government signed an MoU with the IAEA for assistance in the implementation of a nuclear program according to the Milestones approach.1347

The country remains far from fulfilling the basic conditions for a nuclear power program. Ecuador has an installed electricity generating capacity of less than 10 GW and it is unclear whether the grid could deal with a nuclear reactor of any size. It has no regulatory framework and no institutions like a nuclear regulator or a nuclear waste management agency.

Estonia

In November 2020, the Estonian government decided to set up an interministerial Nuclear Energy Working Group (NEWG) that would also convene members from other national organizations. The NEWG’s final report explains that “The ultimate goal of the NEWG is to form coordinated views with the public on the possibilities of introducing nuclear energy in Estonia and to submit its conclusions and proposals to the Government of the Republic.” The Group concluded that “the introduction of nuclear energy in Estonia is feasible” and that “small modular reactors (SMRs) with a power capacity of less than 400 MVA (400 MW) would be suitable for Estonia.” It also pointed to a number of significant challenges:

• “The country needs to make an informed decision on the introduction of nuclear energy, based on a broad political consensus. The Riigikogu [Estonian Parliament] can decide whether to support the launch of a nuclear energy programme, which includes the creation of a regulatory and legal framework, the launch of planning processes and the selection of a suitable technology and developer.”
• “…the launch of a nuclear energy programme will require the regulation of more complex activities and the implementation of new measures.”
• “The construction of a nuclear power plant will require developing a wide range of skills, including acquiring knowledge and experience through international cooperation. Emergency planning and security aspects also need to be considered.”1348

In a significant development, in June 2024, the Riigikogu passed a resolution, based on the NEWG’s conclusions, on supporting the adoption of nuclear energy in Estonia which would allow for “the drafting of the Nuclear Energy and Safety Act and, if necessary, amending and supplementing the existing legislation, as well as the establishment of an institution regulating the safe use of nuclear energy and the development of sectoral competences.”1349 Surprisingly, the resolution was passed with only 41 members voting in favor, while 25 voted against and two abstained, although 55 members had submitted the resolution.

Estonia’s Prime Minister appears well aware of the long lead times of nuclear projects. In April 2025, he stated in response to parliamentary questions that “the construction of nuclear energy is a long process, even if it’s very fast-tracked – it’s an approximate estimate… it takes about 10 to 15 years” pointing to the opportunities of renewables “in the meantime”.1350

Indonesia

The Indonesian government has been contemplating the launch of a national nuclear program for many years with little concrete commitments so far. In its statement to the IAEA General Assembly in September 2024, the Indonesian statement said:1351

We are developing a comprehensive Nuclear Power Plants (NPPs) roadmap, using the INPRO [International Project on Innovative Nuclear Reactors and Fuel Cycles] method under the IAEA’s expertise. (…) We plan to reduce our reliance on fossil fuels and integrate nuclear energy into our energy mix, to generate reliable power and cut greenhouse gas emissions for a more sustainable future.

A senior official at the Indonesian Presidential Office told Reuters in May 2025 that the country is planning national power generating capacity of over 100 GW by 2040, including 75 GW of renewables, 18 GW of natural gas, and 10 GW of nuclear. According to the official, Rosatom, China National Nuclear Corporation (CNNC), Rolls-Royce, EDF, and NuScale “have shown interest in Indonesia’s nuclear power ambitions.”1352

It seems like a long way yet to the implementation of a nuclear construction project in Indonesia.

Jordan

Jordan officialized its interest in nuclear power by establishing the Jordan Atomic Energy Commission (JAEC) through law in 2008.1353 Since then, JAEC has announced plans to import diverse nuclear reactors. The most advanced of these plans was a 2015 agreement with Russia to build two 1000-MW nuclear reactors.1354 But this was ultimately canceled in 2018.1355 Since then JAEC has expressed interest in a wide variety of Small Modular Reactor (SMR) designs developed by companies based in different countries. Indeed, by 2024, the JAEC webpage was announcing that the country had signed “more than 15 nuclear cooperation agreements”.1356 In February 2025, JAEC signed a “memorandum of mutual understanding on cooperation in the field of peaceful use of nuclear energy” with Kazakhstan’s Energy Ministry.1357 The envisaged cooperation likely mainly relates to uranium mining and exploration in Jordan, as an MoU with Kazatomprom to that extent was signed at the time,1358 and Kazakhstan has no reactor design to offer.

Jordan has so far not selected a design, vendor, site, or completed a financing package, let alone started building any nuclear power plant, small or large.

Despite these many agreements, Jordan has so far not selected a design, vendor, site, or completed a financing package, let alone started building any nuclear power plant, small or large. This was underscored in July 2024, when a statement by Khaled Al-Khasawneh, Commissioner of Nuclear Reactors at the JAEC, could only mention “plans to incorporate small modular reactors…into the national energy mix” and how “technical assessments and economic feasibility studies for several SMR designs” were “in progress, along with detailed feasibility studies on using nuclear power for water desalination and pumping in Jordan.”1359

In contrast to its nuclear plans, Jordan has been expanding renewable energy: capacity has been growing quite rapidly, from 200 MW in 2015 to 2.7 GW in 2024, which consisted mostly of solar energy (2.1 GW), followed by wind power (0.6 GW), renewable energy representing 38 percent of total electricity capacity.1360 The government is targeting having solar power and other renewable energy sources contributing 50 percent of the total electricity generation by 2033.1361 According to the Ministry of Energy and Mineral Resources, the 2030-target for renewable energy sources will already be met “between 2027 and 2028”, serving as “incentive to raise future ambitions even further.”1362

Kazakhstan

Kazakhstan is one of five countries in the world to have abandoned commercial nuclear power—while currently contemplating to restart a program—the others being Germany, Italy, Lithuania, and Taiwan. Kazakhstan serves as a lens through which the full spectrum of nuclear energy applications—ranging from commercial power production and uranium mining to nuclear weapons testing—with all their ambiguities, can be examined. The country possesses significant uranium resources and has been the world’s leading uranium producer for more than a decade.1363 This is set against the backdrop of Kazakhstan abandoning its commercial nuclear program in 1998, which had been based on a Soviet-era BN-350 fast breeder reactor.1364

The construction of a new power plant was brought to the table in the 1990s after Kazakhstan’s independence, but the plan became more significant in 2024, when a referendum on the future of Kazakhstan’s civil nuclear program was held. Kazakhstan is heavily dependent on fossil fuels that in 2024 sourced 85 percent of its electricity generation.1365 In recent years the country has faced an increase in electricity consumption with production struggling to keep up.1366

In April 2019, during a visit by Kazakhstan’s then-acting President Kassym-Jomart Tokayev to Moscow, Vladimir Putin raised the possibility of constructing a new nuclear power plant in Kazakhstan with the support of Russian State Atomic Energy Corporation Rosatom. Government sources reportedly commented that while no decision had been made, a potential site had been identified.1367 However, later that same year, shortly after his election as president, Tokayev reportedly denied any immediate plans for building a new nuclear plant in Kazakhstan, while stating that if such a project were to proceed, a referendum would be held.1368

In February 2022, the government presented its “Green with NPP” scenario, which would focus on “green energy sources, gas and nuclear generation” to commission 17.5 GW of new capacity by 2035.1369 At the time, it was reported that the government had been considering six suppliers for SMRs or large reactors as part of its ongoing feasibility studies: Korea Hydro & Nuclear Power (KHNP), NuScale, GE Hitachi, China National Nuclear Corporation (CNNC), Rosatom, and Électricité de France (EDF). But in June 2022, NuScale and GE Hitachi were excluded from the process as their proposed technologies had not been implemented anywhere yet.1370

In September 2023, Kazakhstan’s President Kassym-Jomart Tokayev suggested holding a national referendum in 2024 on the question of building nuclear power reactors. But according to media reports, in May 2024, a spokesperson of the Energy Ministry stated that the referendum had been delayed “over fears that the Kazakh population could vote against the plant’s construction.”1371 Ahead of the referendum, the government had indeed strongly advocated in favor of the nuclear program, leaving little to no room for alternatives: “As world practice shows, construction of a nuclear power plant seems to be the only true solution to replace retiring capacities and ensure energy independence of Kazakhstan. At the present stage, there is no alternative to nuclear power plants.”1372

In September 2024, the platform AES Kerek Emes (“No Need for a nuclear power plant”) held a panel discussion addressing the question, “Does Kazakhstan need nuclear power?” The critiques raised by participants during the panel discussion covered a range of concerns, including ecological risks, challenges of nuclear waste management, delays in project timelines, and significant financial risks associated with such large-scale initiatives. The financial concerns were particularly heightened due to fears of possible corruption. Additionally, Kazakhstani ecologists voiced strong opposition to the use of nuclear power, citing its potential negative impact on the environment.1373

Right before the referendum, at least 12 activists organizing peaceful protests against the nuclear power projects were detained by the police.1374 Most were released on their own recognizance, while five remained in detention pending their trial for “attempting to organize mass riots”, scheduled for July 2025.1375

Nevertheless, the referendum was held on 7 October 2024. Over 71 percent of the votes backed the exploitation of nuclear energy and construction of nuclear power plants in Kazakhstan.1376 Following the referendum, President Tokayev announced the plan to not only accelerate construction of a first plant, but also to create a ‘nuclear cluster’ and fully develop the industry.1377 Further, upon establishing the national Nuclear Energy Agency by presidential decree1378 in March 2025, he stated:

The question is not to ensure only the current energy needs of the country. It is strategically important for us to create a new energy industry that will provide a solid basis for dynamic economic development for decades to come. That is why I consider it necessary to build not one, but three nuclear power plants and, ultimately, form a full-fledged nuclear cluster. Given the exceptional importance of this task, I decided to create a Nuclear Energy Agency under the President.1379

In February 2025, by decree, the site of the first nuclear power plant was officially narrowed down to the Zhambyl district of the Almaty region.1380 More specifically, Ülken, a small settlement at the southern end of Lake Balkhash, has been mentioned as the preferred location for the plant.1381 Although the village was not officially named in the government decree, civil activist Alnur Ilyashev filed a lawsuit challenging the government’s decision to define this location for the plant. Ilyashev raised concerns about the proposed site’s proximity to the Sary-Shagan military ballistic missile defense test site, which is just 22 kilometers from the borders of the Ülken rural district. He argues that this poses significant risks, citing International Atomic Energy Agency (IAEA) safety standards, which recommend a much greater distance between such facilities and sensitive areas.1382

In March 2025, the government planned to select a vendor or consortium by 1 July 20251383 and “conclude an intergovernmental agreement and relevant contracts” by November 2025.1384

Until 14 June 2025, the list of possible companies to build the power plant included Russia’s Rosatom, China’s CNNC, France’s EDF, and Korea’s KHNP. The envisaged designs all corresponded to large units.1385 But finally, the Atomic Energy Agency of Kazakhstan announced Rosatom as the technical supplier of the first new Kazakhstani nuclear power plant.1386 Possibly to dispel the doubts of the critics, Almasadam Satkaliyev, chairman of the Atomic Energy Agency, assured that “the station will be completely Kazakhstani. The owner, operator, and manufacturer of uranium raw materials will be the Republic of Kazakhstan.”1387 On 20 June 2025, both parties agreed on an “Indicative Road Map…envisaging the stages of project preparation and implementation, including surveying, conclusion of an EPC [Engineering, Procurement and Construction] contract and development of project documentation,” while Atomstroyexport and Kazakhstan Nuclear Power Plants concluded a framework agreement.1388 Although no price for the project has been communicated so far, the intergovernmental agreement is reportedly planned to be signed by the end of 2025.1389

On the same day as proclaiming Rosatom as the winning bidder, Chairman Satkaliyev of the Atomic Energy Agency issued a separate statement setting the stage for a later agreement with CNNC:

We plan to sign a separate general agreement with China on cooperation in the nuclear industry. And we want to see Chinese technologies in Kazakhstan for the construction of another nuclear power plant…Objectively, there are not many countries in the world that can master the entire nuclear cycle on their own. And China is undoubtedly one of the countries that has all the necessary technologies and the entire industrial base, and our next main priority is cooperation with China…We have a very high level of agreement. We are interested in adopting Chinese experience, we understand their ability to carry out construction quickly and efficiently, and we have already begun working in this direction1390

According to the government, work is already underway to “select locations for two additional plants,”1391 among them is the town of Kurchatov in Eastern Kazakhstan.1392

In recent years, Kazakhstan has increasingly engaged in strategic partnerships concerning uranium mining, e.g., with France,1393 and uranium supply deals with countries seeking an alternative to Russian uranium following the Russian invasion of Ukraine.1394

Saudi Arabia

Saudi Arabia established The King Abdullah City for Atomic and Renewable Energy (KA-CARE) in 2010.1395 Progress in the past decade and a half has been slow at best, and mostly involved officials reiterating plans to build nuclear plants. As the Saudi Minister of Energy declared at the General Conference of the IAEA in September 2024, “the Kingdom is moving towards utilizing nuclear energy and its radiation applications for peaceful purposes… including the construction of the first nuclear power plant in the Kingdom.”1396

The slow rate of progress is also seen in choosing the potential builders of Saudi Arabia’s first nuclear plant. In mid-2022, KA-CARE invited bids from South Korea, China, France, and Russia to construct two nuclear reactors.1397 Since then the deadlines have been extended several times. One unnamed source told the Middle East Economic Digest (MEED) in early 2025 that the bid deadline is “more like a moving target, running in parallel with the progress in the bilateral government-to-government talks.”1398

In 2023, Nuclear Intelligence Weekly reported that Saudi Arabia had received bids from Korea Electric Power Company (KEPCO), China National Nuclear Corporation (CNNC), Russia’s state-owned Rosatom, and France’s EDF.1399 However, in May 2025, KEPCO reportedly said that it has been in talks with Saudi Arabia and the company is “pushing to submit bids to six different energy projects in the region, including a nuclear power plant construction project and a combined heat and power plant project in Saudi Arabia, this year.”1400 This suggests that it has not yet put in a bid.

No U.S. company has put in a bid. The Congressional Research Service says as of September 2024 that

Congress has directed that no funds appropriated for the Department of State, foreign operations, and related programs ‘should’ be used by the Export-Import bank to support nuclear exports to Saudi Arabia until the kingdom has a 123 agreement ‘in effect’; ‘has committed to renounce uranium enrichment and reprocessing on its territory under that agreement’; and has ‘signed and implemented’ an Additional Protocol with the IAEA.1401

However, in April 2025, U.S. Secretary of Energy Chris Wright reportedly declared that his government “has revived talks with Saudi officials over a deal that would give Saudi Arabia access to U.S. nuclear technology and potentially allow it to enrich uranium” that is meant to “enable the kingdom to develop a commercial nuclear power industry.”1402 The following month, upon the conclusion of preliminary agreements on energy and critical minerals cooperation, the U.S. Department of Energy specified: “The two sides also outlined areas for cooperation on civil nuclear energy, including safety, security, and nonproliferation programs; vocational training and workforce development; U.S. Generation III+ advanced large reactor technologies and small modular reactors; uranium exploration, mining, and milling; and safe and secure nuclear waste disposal.”1403 Even if these talks move forward, it is unlikely to result in a reactor export in the short term.

In the meanwhile, total renewable energy capacity in Saudi Arabia has grown from 24 MW (solar only) in 2015 and 3 GW in 2023 to 4.7 GW in 2024, a 200-fold increase over a decade.1404 Renewable energy still constitutes only 5 percent of the total electricity capacity. The bulk of renewable capacity remains solar with over 90 percent. Wind energy capacity has been stagnant at 403 MW since 2022. Production of electricity from solar and wind power plants in 2024 amounted to only 8.2 TWh and 1.6 TWh respectively, only 2.2 percent of the total amount electricity generated, the rest being covered by natural gas (over 63 percent) and oil (34.5 percent).1405

Uzbekistan

Although Uzbekistan’s uranium reserves are smaller than those of neighboring Kazakhstan, the country has significantly increased its uranium mining activities in recent years.1406 Similar to Kazakhstan, electricity production in Uzbekistan is heavily dependent on fossil fuels. However, unlike Kazakhstan, where coal is the main energy source, Uzbekistan’s main source of power generation is natural gas, accounting for over 80 percent of the electricity mix in 2023.1407 As in Kazakhstan, the demand for energy in Uzbekistan is projected to double by 2030.1408

In 2017, Uzbekistan entered into a framework agreement on nuclear cooperation with Russia. This was followed in September 2018 by a more specific agreement for Rosatom to construct two VVER-1200 reactors.1409

In an April 2019 interview with Nuclear Engineering International (NEI), Jurabek Mirzamakhmudov, Director General of Uzatom—the state-controlled Atomic Energy Agency of Uzbekistan—announced plans to conduct site analysis at three potential locations. Mirzamakhmudov stated that the project would be partially financed through a soft loan from Russia. While the reactors are intended primarily to supply power for domestic use, some of the electricity could also be exported to neighboring countries, such as Afghanistan.1410 It was later reported that the goal was to select and license a site by September 2020;1411 however, this timeline was not met.

The first export contract for an SMR anywhere in the world

In May 2022, Mirzamakhmudov announced that a site had been selected in the Farish district of the Jizzakh region, near Lake Tuzkan, to host two Rosatom supplied VVER-1200 reactors. He noted in an interview that while the financing package was still under negotiation, sanctions imposed on Russia due to the war in Ukraine would not affect the project.1412

Nevertheless in May 2024, Uzbekistan apparently moved away from the idea to build two large nuclear reactors and signed an agreement with Russia’s Rosatom to build six 55-MW Small Modular Reactors (SMRs) in the eastern Jizzakh region, which marks the first export contract for an SMR anywhere in the world.1413 The design to be implemented is the RITM-200N, an adaptation for land-based applications of the RITM-200 design that is used on the Russian icebreaker fleet.1414 If built, this will be the reactor’s first installation outside Russia.

In April 2025, per Rosatom “on-site work began to set up a construction yard that will accommodate production facilities (particularly prefabrication workshops), administrative buildings, and warehouses” to support the construction of the RITM-200 cluster in the Jizzakh region.1415 Additionally, Rosatom began with the production of the reactor pressure vessel of the first reactor block of the RITM-200 cluster.1416 In April 2025, the head of Rosatom was quoted by Russian news agency TASS as saying that licensing and first concrete would be “implemented in a year or maybe earlier.”1417

In April 2025 Shanghai Electric was named as potential manufacturer for the turbines of the reactor following a meeting between Uzatom, Rosatom, and Shanghai Electric at the office of China National Nuclear Corporation (CNNC).1418

In June 2025, Deputy Prime Minister of the Russian Federation, Alexander Novak, reportedly stated that “Today, specific economic parameters are being discussed. Two units of 1000 MW each and two units of 55 MW each.”1419 The next day, Rosatom announced that an agreement with Uzatom had been officialized “to study the possibilities of implementing a project to construct a large nuclear power plant in the Republic of Uzbekistan” which “specifies the main terms and conditions for the potential implementation of the construction of two VVER-1000 power units with possible upgrade to four units.”1420 Rosatom had reportedly already signaled in April 2025, that “it is not a question of if, but when the leadership of Uzbekistan returns to the topic of a large nuclear power plant, we will have proposals ready.”

This statement contradicts both the initial plan to build two VVER-1200 reactors and the official plan to build six RITM-200N reactors.

Thus, according to some sources the plan to build six RITM-200N reactors was “upgraded” to include two VVER-1000 reactors and two RITM-200N reactors. It has not yet been disclosed whether the VVER-1000 and RITM-200N reactors will be built at the same location or at separate sites.1421

Russia Nuclear Interdependencies

Russia is a major global supplier of nuclear fuel services, including uranium mining, conversion, enrichment, and fuel assembly fabrication for Soviet-designed VVER pressurized water reactors, of which there are 19 in the E.U. and 15 in Ukraine. In particular, since Russia’s full-scale invasion of Ukraine in February 2022, E.U. members and others in the region have discussed and taken measures to deprive Russia of the considerable revenue streams provided by such businesses and reduce the inherent risk of the region’s dependence on Russia. But compared to Russian supplies of oil, natural gas, and coal, the nuclear sector has received less attention. While the U.S. banned the import of uranium products from Russia in May 2024 (allowing for conditional waivers until 2028), the E.U. has not established any sanctions in the nuclear sector—a strong indicator of dependency on Russia.

However, as highlighted in this chapter, the dependency is multi-facetted; Russia is also dependent on the West for European companies’ cutting-edge technology used in its reactors, and business interests of these same companies have certainly contributed to prevent or delay sanctions.

This chapter provides an updated overview of these interdependencies, with a special focus on the supply of VVER fuel elements in the E.U., which in the past have exclusively been provided by Russia. Meanwhile, all E.U. and Ukraine operators of VVER reactors have signed contracts with alternative fuel suppliers from Western countries, where the Russian government-controlled company, Rosatom, is still partially involved. Further, the E.U. announced a Roadmap Towards Ending Russian Energy Imports that is to be translated into legislative proposals, which will possibly include measures in the nuclear sector. Western companies have also announced significant expansions of their uranium enrichment capacities to accelerate the end of Russian dominance in this area.

Russia’s Role in the Global Nuclear Fuel Supply Chain

According to the World Nuclear Association (WNA), as of 2022, Russia held about 20 percent of the world’s estimated primary uranium conversion capacities1422 and 44 percent of enrichment capacities globally.1423 Russia therefore clearly dominates uranium enrichment services globally. Its share of the worldwide capacities will change over the coming years due to added capacities in other countries, but a clear picture on the global scale is yet to emerge—beyond the E.U. only partial information is available on individual countries’ uranium conversion and enrichment capacities. In 2022, Russia also contributed 5 percent to the world’s natural uranium production, while Kazakhstan provided 43 percent and Uzbekistan an estimated 6.7 percent (both countries have close ties to Russia).1424

In 2024, the U.S. imported 335 tons of enriched uranium from Russia (valued at US$624 million), down from 702 tons (US$1.2 billion) in 2023—the highest volume since 2013—and 588 tons (US$830 million) in 2022. Thus the U.S. reduced the quantity of these imports last year by more than 52 percent compared to 2023 when Russia was its largest supplier.1425 In May 2024, the U.S. introduced a ban on the import of Russian uranium products into the U.S. starting in August 2024, but it allows for a waiver process through 1 January 2028.1426 Russia responded with a “tit-for-tat” ban on enriched uranium exports to the U.S. in November 2024. However, a Kremlin spokesperson reportedly stated that “in cases where it serves our interests, Russia’s Federal Service for Technical and Export Control may decide to exclude certain items from this list of bans.”1427 In other words, the Russian ban is supposed to work exactly as the U.S. one: if in the national interest, a waiver may be granted. Whether the reduced U.S. imports of enriched uranium in 2024 reflect an effect of these mutual sanctions or not remains unclear.

On the other hand, an analysis of Chinese trade data shows increased deliveries of enriched uranium from Russia to China and a significant increase in deliveries from China to the U.S. in 2023 and 2024.1428 In particular, China’s imports from Russia reached a maximum of 748 t in 2024,1429 after 467 t in 2023, 685 t in 2022, 0 t in 2021, 48 t in 2020, and 504 t in 2019. In the same time frame, U.S. imports from China were 7 t in 2019, no reported imports between 2020 and 2022, but 294 t in 2023 and 124 t in 2024, so that China ranked as the sixth largest enriched uranium supplier (of ten) to the U.S. in 2023 and 2024.1430 Whether these deliveries are to be considered as a “displacement” of the reduced deliveries from Russia to the U.S. cannot be established from this data.

Figure 54 shows the origin of natural uranium, conversion services, and enrichment services supplied to the E.U. based on data provided by the Euratom Supply Agency (ESA). After a spike in all three categories in 2023, Rosatom’s share decreased in 2024, but it still provided 16–24 percent of the respective services in 2024 compared to a 23.5–38 percent share in 2023.

1. Providers of Nuclear Fuel Services to the E.U.

Source: ESA, 2022–2025

Note: Total uranium delivered to E.U. utilities in 2021 also includes 196 t (1.6 percent) of re-enriched uranium, not represented.

According to ESA statements in early 2023, without Russia the nameplate capacity of uranium conversion and enrichment plants in the E.U. would be “sufficient for the E.U. to be self-dependent,” but the “Global West” would be missing enrichment capacity of 3,500–8,000 tSWU (thousand Separative Work Units).1431 According to their respective 2024 annual reports, Urenco (U.K.-Netherlands-Germany) and Orano (France) have announced plans to expand their global capacities by 1,800 tSWU1432 and 2,500 tSWU,1433 respectively, i.e., 4,300 tSWU in total, accounting for a little more than the minimum capacity gap of 3,500 tSWU claimed earlier by ESA. The “Global West”, therefore, would significantly reduce its dependency on Russia for enrichment capacity.

The E.U.’s fuel assembly imports from Russia for the 19 Soviet-designed VVERs (Vodo-Vodianoï Energuetitcheski Reaktor) have declined to 438 tons in 2024 from a peak of 573 tons in 2023, though they still exceed the 314 tons in 2022 (see Figure 54). The high volumes in 2023 and 2024 most likely reflect stockpiling, given political uncertainties and anticipated sanctions due to Russia’s war in Ukraine, until alternative fuel supplies become available.

1. E.U. Imports of Russian Nuclear Fuel Elements

Source: Eurostat Database, 20251434

Note: EU Total does not count imports to Bulgaria (data not available). For more detail, see investigation by Bellona.1435

While the E.U.’s sanctions against Russia in response to its invasion of Ukraine—the 17th package of which was approved on 20 May 2025—have excluded the nuclear sector thus far, on 6 May 2025 the E.U. announced a Roadmap Towards Ending Russian Energy Imports, including nuclear supplies. The proposed measures include the following:

• trade measures to make the purchase of Russian enriched uranium less attractive
• a ban on new contracts with Russian suppliers through the Euratom Supply Agency
• development of a step-by-step phaseout plan for Russian nuclear imports in cooperation with Member States by the end of 2025

However, Hungary and Slovakia immediately rejected the plan, with the Slovakian prime minister qualifying the scope of measures as “simply economic suicide”.1436

Five E.U. countries—Bulgaria, the Czech Republic, Finland, Hungary, and Slovakia—operate 19 Soviet-designed VVER reactors, 15 of which are 440s and four are 1000s. These countries are dependent on the supply of specific fuel and, in parallel, have particularly high shares of nuclear power in their respective electricity mixes, ranging from around 40 percent to over 60 percent (see Table 16). From the start of their reactors’ operations, these countries have almost exclusively received fuel assemblies from TVEL, a 100-percent subsidiary of Rosatom.

1. Fuel Supply for Soviet-designed Reactors in the E.U. and Ukraine (as of mid-2025)

Sources: Various, compiled by WNISR, 2025; all available in Annex 3.

Table 16 provides an overview of publicly available information on fuel supply agreements for the VVERs in the E.U. The multiple diversification initiatives aiming at the phaseout of fuel deliveries from Russia are described in more detail in the respective country sections. All VVER-operating countries in the E.U. have agreements in place for alternative fuel supply as of mid-2025, including Hungary which signed an agreement with Framatome in October 2024 but which remains the only country without a Westinghouse contract. On the other hand, to date, neither Finland nor Ukraine has signed an agreement with Framatome. In November 2024, Slovakia extended its supply contract with TVEL until 2030.1437

Diversifying fuel supply for these reactors is complex. Westinghouse started manufacturing VVER-1000 fuel assemblies in 1998 and supplied Czech reactors at Temelín in their first decade of operation, starting in 2000.1438 Westinghouse has since become the sole supplier in Ukraine and, starting in 2023, also provided first loads to VVER-440 reactors in the Czech Republic, Finland, and Ukraine.1439 Development of Westinghouse fuel is supported by a Euratom funded project named APIS (Accelerated Program for Implementation of secure VVER fuel Supply) that started in 2023 and runs until 31 December 2025.1440 In 2023 Ukrainian energy company Energoatom announced its intention to manufacture nuclear fuel for its reactors in cooperation with Westinghouse.1441 In May 2025, Energoatom received a license from Westinghouse for manufacturing fuel assembly components and announced that the first Ukrainian fuel assembly from its own production line will be loaded into a reactor core in 2027.1442 In the long term, Energoatom plans to manufacture half of Ukrainian nuclear fuel, the other half being provided by Westinghouse. The difficulty and complexity of diversifying nuclear fuel supply may be illustrated by the fact that Finland, on one hand, cancelled the Hanhikivi reactor project with Rosatom in 2022, soon after Russia’s invasion of Ukraine, on the other hand, it almost doubled its import of Russian nuclear fuel in 2024 (see Figure 55), despite a fuel supply contract with Westinghouse in place since 2022. Simply switching the supplier, therefore, does not seem possible in the short term as it can be assumed that Finland is eager to become independent of Russia.

In addition to Westinghouse, Framatome is expected to join the ranks of VVER fuel manufacturers, following a dual track approach for short-term and long-term solutions. In the short-term, Framatome intends to manufacture the existing Russian VVER fuel design under TVEL license. To this end, Framatome has established a joint venture with TVEL for manufacturing VVER fuel at its plant in Lingen, Germany. This plan has drawn widespread criticism and is pending because, as of mid-2025, it has not been approved by the German government (see the following section for details). If the government rejects the plan, Framatome will not be able to fulfill its supply contracts, for example, with the Czech Republic which should have started already in 2024 and with Bulgaria which should start in 2025, as can be seen from Table 16.

Framatome has been progressing on its own long-term VVER fuel design since 2018. Framatome hopes to begin fabricating the first lead-test assemblies in 2026, and after a test phase, the technology could be available in the early 2030s, according to the head of Fuel Business at Framatome.1443 In addition, Euratom project SAVE (Safe and Alternative VVER European Project), led by Framatome, gathers seventeen partners from seven E.U. Member States as well as Ukraine to develop the new, European VVER fuel design. The project started in June 2024 and will run until 30 June 2028.1444

Framatome and the Lingen VVER Fuel Manufacturing Plant Project1445

In contrast to Westinghouse, which has developed and manufactured fuel assemblies for VVER-type reactors at least since 1998, Framatome has been completely absent from this market. However, on 2 December 2021, twelve weeks before Russia invaded Ukraine, Framatome announced that it had signed a long-term strategic agreement with Rosatom aimed at “further expanding the companies’ efforts to develop fuel fabrication and instrumentation and control (I&C) technologies.”1446 Regarding the joint “efforts to develop fuel fabrication”, Framatome and TVEL had already founded a joint company in France in 2019 which was later renamed to “European Hexagonal Fuel SAS”.1447 In March 2022, Framatome’s subsidiary Advanced Nuclear Fuels (ANF) filed an application with the environment ministry in Lower Saxony, Germany seeking permission to modify ANF’s facilities in Lingen to enable the licensed manufacturing of VVER fuel assemblies.1448

Following German atomic law, ANF’s request was subject to a public inquiry in early 2024 resulting in more than 11,000 comments.1449 To consider these comments, the Lower Saxony Government held a public hearing in Lingen in November 2024. The main objections concerned TVEL’s role in the manufacturing, raising issues of potential espionage risks and threats to national and energy security.1450 As of mid-2025, no decision had been issued by the Lower Saxony Government. The environment minister, a Green Party member, is opposed to the project,1451 but he can only examine the application based on his mandate under federal atomic law and does not have a veto right. It is up to the federal government to approve or block the initiative.

ANF nevertheless started preparations for manufacturing VVER fuel assemblies in Lingen. In May 2024, ANF confirmed that production machinery had been set up at a separate facility outside the Lingen factory, and that it had been tested in April 2024.1452

The apparent lack or delay of ‘real’ diversification away from Russia has raised the question why Framatome partnered with TVEL to enter the VVER fuel market, instead of Westinghouse which has developed and delivered VVER fuel assemblies for nearly three decades, and why it stuck to this decision even after February 2022, particularly given Rosatom’s proactive role in the Ukraine war.1453 An answer may lie in the assumption that Framatome regarded Westinghouse as its main competitor, so that after the start of the Ukraine war, Framatome, without a short-term solution, would have been overtaken by Westinghouse until having developed its own fuel, which would likely happen after 2030. Framatome therefore needed Rosatom to enter the market for VVER fuel in a timely manner.1454

Another question is why TVEL entered into the joint venture with Framatome even before the Ukraine war because it effectively meant losing market share and revenues. Some hints might be contained in the overall strategic cooperation agreement signed between Framatome and Rosatom in December 2021 which aims, among other things, “to develop fuel fabrication and instrumentation and control (I&C) technologies.”1455 Thus, Rosatom and Framatome’s cooperation covers far more than fuel manufacturing. Possibly in return for sacrificing market shares in VVER fuel manufacturing, Rosatom could have gained further access to Framatome’s I&C or other technologies, while Framatome could enter the VVER fuel market and extend its business with I&C technology. This may also partially answer the earlier question as to why Framatome stood by its decision to undertake a joint venture with TVEL even after February 2022: strong mutual business interests between Framatome and Rosatom.

Russia’s Dependencies and Potential Sanctions

The long-standing business relations between Framatome, its partner Siemens Energy, and Rosatom create a mutual dependency, including significant Russian dependence on the West which may get increasingly relevant when further sanctions are considered. I&C technology discussed in the preceding paragraph is “the brain and central nervous system” of a power plant, according to Framatome.1456 For a decade Russia has received from Framatome and Siemens Energy (spun off from Siemens in April 2020) I&C technology not only for the initial construction but also modernization and servicing of existing reactors which last long into the future. The most recent reported case involves a Framatome I&C system, due to be fully installed at the newbuild Kursk II nuclear power plant (two units under construction) by the end of 2025.1457

The I&C systems provided by Framatome and Siemens Energy cover different aspects of the operation of a nuclear power plant so that the companies are not competitors but partners.

Framatome offers Teleperm XS (where the S stands for security), i.e., a security system, also called reactor protection system, performing automatization functions relevant to security.1458 It is based on the digital automation system Teleperm which was developed by Siemens in the 1980s. In 2000, Framatome and Siemens formed a joint venture for nuclear technology (in which Siemens had a minority share of 34 percent) that was later named AREVA NP. After Siemens announced its exit from AREVA NP in 2009, Teleperm XS became a product of AREVA NP alone and later of AREVA NP’s successor Framatome.

The I&C system offered by Siemens Energy, on the other hand, is SPPA-T2000/T3000 (earlier named Teleperm XP), a process control system for a power plant’s operation, including monitoring and control of the whole plant in a control room.1459

Framatome and Siemens Energy have contracted their respective I&C systems as a consortium to Rosatom for the Paks II project in Hungary (see section on Hungary) as well as for a range of other Russian reactor projects around the world, including in Russia itself, e.g., for the new build projects Novovoronezh 2-1 and 2-2, Leningrad 2-1 and 2-2, and the modernization of Kola-3 and -4 that were carried out between 2009 and 2016 together with Framatome’s predecessor AREVA NP.1460 Since the I&C systems are deeply engrained into the operation of these reactors, their continuous service and modernization needs create a strong, long-lasting Russian dependency on its western partners.

Siemens Energy terminated their business with Russia after the start of the Ukraine war and claimed in 2023 that their turnover with Rosatom was “in the one-digit million range” in 2022.1461 In January 2025, a Siemens Energy spokesperson reportedly stated that “Siemens Energy ended all activities in Russia earlier and no longer has any contractual relationships there” and “now only has to fulfill older, outstanding contracts, concluded before the start of the war in Ukraine.”1462 This statement could be interpreted to mean that Siemens Energy still provides service and modernization of I&C systems of reactors in Russia “to fulfill older, outstanding contracts” which apparently would violate E.U. sanctions. In response to corresponding questions from WNISR, a Siemens Energy representative responded that the company would not provide services to reactors in Russia and therefore also would not circumvent sanctions.1463

However, Siemens Energy’s operations in Moscow are apparently continued by the company Energy Management LLC, which is run by former Siemens manager Oleg Titov1464 and claims on its website to “continue the traditions of Siemens.”1465 To perform its operations, this company needs spare parts, new parts, new software versions etc., which earlier were delivered by Siemens Energy. Meanwhile, these products are allegedly offered by other companies via other channels. For example, the company Industrial Technologies LLC in Sverdlovsk, Russia, runs a website “Siemens.B2B”, which offers to provide Siemens products as an “an official dealer of SIEMENS in Russia” and claims to have its “own logistics solutions for the supply of products to the Russian Federation, limited or prohibited for sale through the sanctions policy of European manufacturers.”1466 

In response to corresponding questions from WNISR, Siemens Energy declared to have “no ongoing transactions or other business activities with the two companies mentioned”, (i.e., Energy Management LLC and Industrial Technologies LLC) and to have “taken all necessary measures and established processes to comply with the relevant provisions of German, European, and US (re-)export control law”, including “careful checks and measures to protect against the circumvention of sanctions”, wherein “[e]very order is checked independently of the country of destination based on the applicable goods lists, the final destination, and the intended use of the goods.”1467

In the past decade, Russia turned into the world’s dominant international reactor technology supplier. With the exception of one Chinese project in Pakistan and one “first concrete” in South Korea, the Russian industry implemented all other 13 construction starts in the world outside China between December 2019, when EDF officially began construction at the U.K.’s Hinkley Point C Unit 2, and mid-2025 (see Overview of Current Newbuild).

A major participant in Rosatom’s reactor projects is Arabelle Solutions (formerly known as GEAST, for GE-Alstom), the Arabelle turbine manufacturer in France. In 2024, Framatome’s majority owner, EDF, acquired the company from GE1468 after a lengthy process of over two years.1469 The company produces turbines for nuclear power plants and is highly dependent on the niche market virtually entirely controlled by the Russian nuclear industry over the past five and a half years. Reportedly, Rosatom represented about half of Arabelle Solutions’ turnover as of 2022,1470 implemented through a joint venture with Rosatom in Russia, Turbine Technologies AAEM LLC, which supplies turbine islands for Rosatom reactor projects (see box hereunder). It was therefore no surprise that just before Russia’s invasion of Ukraine the French government had offered to sell Rosatom a 20-percent share in the company.1471

Turbine Technologies AAEM LLC – EDF’s Outpost in Russia

In 2007, French industrial giant Alstom and Rosatom subsidiary Atomenergomash JSC (AEM) formed the joint venture Alstom Atomenergomash LLC (shortened as AAEM), in which Alstom owned 49 percent and Atomenergomash 51 percent.1472 The company was later renamed to “Turbine Technology AAEM LLC” and is headquartered in St. Petersburg, Russia.1473 After several transitions involving the U.S. company General Electric (GE), Alstom’s share is now owned by France’s Arabelle Solutions, which in 2024 was acquired by the French utility (and Framatome owner) EDF.1474 The purpose of AAEM is to provide turbine hall equipment of VVER reactors—the largest part of a nuclear reactor’s equipment except for the nuclear island itself—based on the Arabelle turbine. The manufacturer considers his steam turbine to be best-in-class with regard to efficiency and reliability,1475 and therefore provides a significant competitive advantage. AAEM will supply the turbine islands for Rosatom reactor projects including four at Akkuyu in Türkiye, four at El Dabaa in Egypt, and two at Paks II in Hungary.1476 Therein, Arabelle Solutions will provide the turbines, GIGATOP four-pole generators, and condenser vacuum systems, while pumps, heat-exchange, and other auxiliary equipment will be manufactured by Atomenergomash and its subsidiaries in Russia.1477 Earlier, AAEM won contracts for reactor projects in Kaliningrad (Russia), Belene (Bulgaria), and Hanhikivi (Finland), but those were cancelled.

Arabelle Solutions’ “heavy order book with Rosatom”1478 for projects with AAEM amounted to €1.36 billion (US$20231.5 billion) at the end of 2023, wherein the revenue already recorded on these projects was €571 million (US$2023617 million), according to the company’s auditor report.1479 This figure is to be compared with Arabelle Solutions’ turnover of €437 million (US$2023472.5 million) in 2023.1480

The relevance of the Arabelle turbines for Rosatom’s business is highlighted by Norwegian think tank Bellona’s study which analyzed the drop in electricity generation in Russia in 2023. The study identified that a major reason for this—based on publicly available information—were defects of Russian made turbines, notably in new reactors such as Leningrad 2-1. Bellona notes that compared to these Russian turbines, Rosatom offers a “more competitive product”, namely the Arabelle turbine, in its foreign projects (in particular in Hungary, Türkiye, and Egypt), which increases the chances of winning tenders. Bellona concluded:

If the West is seriously concerned about how to limit Rosatom’s activity in third countries, then reducing cooperation in the supply of turbine units for new projects could be a serious and significant step.1481

Small Modular Reactors (SMRS)

The Merriam-Webster dictionary defines the term “Potemkin village” as “an impressive facade or show designed to hide an undesirable fact or condition.”1482 The state of Small Modular Reactors (SMRs) today might well be described as a Potemkin Village. Even as the evidence for the high costs and the long timelines for potential future construction becomes clearer, and what is most on display are numerous announcements about future SMRs, usually held out for some time in the 2030s, the industry, politicians, investors, and, last but not least, the media continue to portray SMRs as an indispensable and sure way to solve the climate emergency crisis, and more recently, as a sound way to provide energy for proliferating data centers and demand from generative AI (Artificial Intelligence).

One reason why these claims become more widely believed is the continued financial and political support offered to SMRs by multiple governments, some venture capitalists, and the nuclear industry itself. The OECD’s Nuclear Energy Agency (NEA) reports that “private capital is playing an increasingly important role [in SMR financing], often complementing public matching grants”, and there is approximately “USD2023 15.4 billion of financing towards SMRs worldwide”; of this, roughly “USD2023 10 billion has come from public sources, alongside approximately USD2023 5.4 billion” from private sources.1483 But the NEA has also “identified 127 SMR designs around the world” of which 74 are included in the third edition of its SMR Dashboard. In other words, the US$202315.4 billion of funding is spread out thinly, and most designs have insufficient resources to be developed into ones that can be licensed and built. To put this amount into perspective, as of 2023, just one SMR company, NuScale, years away from construction, had already invested US$1.8 billion (of various origins, including a big portion of public funds) on research and development.1484

This chapter offers brief updates on programs in all those countries developing their own SMR designs (for details, see respective individual Focus Countries sections or dedicated sections in Annex 1) This overview of SMRs in different countries shows that the reality about SMRs is far from the vision portrayed widely. In Argentina, for example, the SMR design that dates back to the 1980s and has been “under construction” since 2014, was officially abandoned in 2024. The SMR design that was to be the first to be built in Canada is no longer under consideration; indeed, the parent company filed for bankruptcy protection in 2024.

Argentina

Argentina’s CAREM (Central Argentina de Elementos Modulares) continues to be the oldest SMR design that actually started building. Under development by the National Atomic Energy Commission (CNEA) since the 1980s,1485 at construction start in 2014, the 25-MW CAREM was “scheduled to begin cold testing in 2016 and receive its first fuel load in the second half of 2017”.1486 As with all other SMR designs so far, that deadline was not met.

In September 2024, due to severe budget cuts, 153 construction workers involved in the CAREM project were reportedly laid off, bringing the total number of layoffs that year till that point to 470.1487 The report went on to state that there were “only 160 builders still in their position” and that “only 30 will remain by November [2024].” The new CNEA president, Germán Guido Lavalle, declared the end of the CAREM project in late 2024. CAREM would not fit into the “commercially viable” category, as he reportedly said in an internal statement “let’s be intellectually honest, we are not going to sell 50 CAREMs, we know that…this reactor is not economically competitive. You only have to stand in front of the construction site to realize that this is not a small modular reactor.”1488

Then, in December 2024, Argentina’s President Javier Milei announced that Argentina aimed to be a pioneer in nuclear power’s “triumphant return”.1489 Reportedly, this involved building an SMR at the existing Atucha site. But rather than CAREM, the reactor to be built would involve “a new conceptual design for a 300-MW SMR developed by Invap”, a company based in Argentina.1490

More details were revealed on 30 May 2025. At an event organized by the CNEA, the chairman of the board of Nucleoeléctrica stated:

The ACR-300, a 300 MW technological marvel designed by Argentine engineers, is a cornerstone of the Nuclear Plan that will position our country at the forefront of the new energy revolution…We are going to begin construction of four modules at the Atucha site, which will nearly double the country’s installed nuclear capacity. Next, we will license this technology worldwide. This will not only transform our energy matrix – it will also reshape Argentina’s export profile.1491

INVAP, in the meanwhile, has obtained a patent from the U.S. Patent and Trademark Office, on the ACR-300 which it initially filed in 2018.1492

If Argentina were to move forward on the ACR-300, it would be starting from a much earlier point in the development cycle for an SMR. The timeline projected by officials—building the first ACR-300 by 2030—is highly unlikely to materialize, especially given the long history of failed promises with the CAREM design. Indeed, even Adriana Serquis, former CNEA president reportedly argued that “this kind of announcement of the Argentine government is to say something for people [to] cheer” because the ACR-300 “has no engineering detail of any kind.”1493

Canada

Government entities in Canada have continued to promote and support small modular reactors, as has been the case since the publication of the 2018-SMR Roadmap.1494 The province of Ontario is the main source of support for what might be Canada’s first SMR construction project. Ontario Power Generation (OPG) is planning for four BWRX-300 reactors designed by GE-Hitachi at the Darlington site.1495 However, in its October 2022 application to the Canadian Nuclear Safety Commission (CNSC), OPG only asked for a license to construct a single BWRX-300 unit.1496 This was approved in April 2025.1497 According to OPG’s website the goal is to “complete construction of the first SMR by the end of this decade, and connect to the grid by the end of 2030.”1498 GE-Vernova’s annual report, released in February 2025, stated that commissioning of the first unit was scheduled for 2029.1499 In 2021, OPG projected it to “be completed as early as 2028”.1500

In May 2025, the Ontario government approved OPG’s plan to spend CAD7.7 billion (US$5.6 billion) to construct the first BWRX-300 unit, which includes CAD6.1 billion (US$4.4 billion) for the actual unit and an additional CAD1.6 billion (US$1.2 billion) “on common infrastructure such as administrative buildings and cooling water tunnels the new reactor will share with three additional BWRX-300s to be built later.”1501 If OPG goes ahead with the remaining three units, the total estimated cost for the project is CAD202420.9 billion (US$202415.2 billion), including “interest charges and contingencies”.

It is not clear if this total includes CAD970 million (US$708 million) provided in 2022 for site preparation by the government’s Canada Infrastructure Bank.1502 The funding announcement stated that it “covers all preparation required prior to nuclear construction, including project design, site preparation, procurement of long lead-time equipment, utility connections, implementation of a digital strategy, and related project management costs.”1503

There is also a SMR engineering and service centre to be built near the Darlington site, for which GE Vernova Hitachi Nuclear Energy “has announced an investment of [CAD]$70 million [US$51 million]”.1504 According to GE-Vernova, the proposed facility will “include a state-of-the-art virtual reality simulator and provide training capabilities to support safe and efficient SMR refueling and maintenance evolutions.”1505

When the CAD970 million announced in 2022 is included, funding for SMRs from various federal and provincial government sources in Canada totals at least CAD1.5 billion (US$1.1 billion) so far.

Among the larger contributions since mid-2024 are CAD80 million (US$58 million) to Saskatchewan government’s Crown Investments Corporation to carry out “pre-engineering work and technical studies, environmental assessments, regulatory studies and community and Indigenous engagement,” CAD55 million (US$40 million) to OPG to “support site planning, site preparation, various procurements and regulatory approvals for the three new SMRs” at the Darlington New Nuclear Project (Units 2, 3, and 4), CAD25 million (US$18 million) to NB Power “for up to 600 megawatts of new small modular reactor (SMR) capacity at the Point Lepreau Nuclear Generation Station”, and CAD13 million (US$9.4 million) to Capital Power Limited Partnership to assess the “potential suitability of three locations in Alberta as potential host locations for SMR deployment.”1506 But there were also large streams of funding for research into various aspects of SMRs, including CAD13.6 million (US$20249.9 million) to eight entities, mostly private companies.1507

The Canadian Nuclear Safety Commission (CNSC) offers an optional three-phase “vendor design review” for nuclear vendors to enable CNSC staff “to provide feedback early on in the design process.” The review “does not involve the issuance of a licence and does not certify a reactor design,” and it “is not required as part of the licensing process for a new NPP [Nuclear Power Plant], and its conclusions do not bind or otherwise influence decisions made by the Commission.”1508

CNSC completed the Phase 2 pre-licensing review of the ARC-100 in early July 2025 that identified “no fundamental barriers to licensing” but it required the company “to perform additional work to address” eight “technical clarifications and findings” if the company or another proponent were to “pursue future VDR [Vendor Design Review] work or licence application reviews”.1509 The CNSC identified some challenges that could pose safety risks if not addressed; for example, the “effectiveness of the proposed means of reactor control and shutdown”. Currently, the only SMR design in the review process is Westinghouse’s eVinci (since June 2023).

The 5-MW Micro Modular Reactor (MMR) design was listed last year as being “on hold”, but is no longer listed.1510 The design’s developer, Ultra Safe Nuclear Corporation (USNC), announced in February 2024 that it was involved in “a reduction of USNC staff and the concentration of efforts on selected markets and customers” because “only a subset” of potential customers for the reactor design “have shown the resolve to incorporate advanced reactors in the near term.”1511 The company’s financial troubles continued since then; in October 2024, USNC filed a Chapter 11 petition for bankruptcy protection.1512 According to one of the documents submitted to the Bankruptcy Court in the State of Delaware, U.S., USNC and its subsidiaries owed over US$16 million to various entities, including US$647,308 to the CNSC.1513

In December 2024, another startup company called Nano Nuclear acquired the patent for the Micro Modular Reactor design and another reactor technology with associated intellectual property rights for US$8.5 million.1514 Although Nano Nuclear has announced that it “is actively preparing to construct a KRONOS demonstration reactor in Canada, where it will enter the licensing process under Canadian Nuclear Safety Commission (CNSC) oversight,”1515 as of July 2025, it had not done so.

Besides Ontario, the other province that has been at the center of SMR activity in Canada is New Brunswick. The province’s electricity company, NB Power, applied to the CNSC in June 2023 for a license to prepare the Point Lepreau site to construct and operate an ARC-100 sodium-cooled fast reactor by the early 2030s.1516 As detailed in earlier WNISR editions, ARC faces financial challenges. In July 2025, following the conclusion of CNSC’s Phase 2 VDR, a company spokesperson announced that the company’s “current focus is on advancing strategic partnership and investment discussions to set the stage for the next phase of design work to support a license to construct application” but when asked by a local newspaper about cost estimates, the spokesperson said “we continue to evaluate the going forward cost estimate through current discussions with strategic partners” and “we are not sharing specific numbers.”1517 But a former ARC CEO put the figure as “likely…between US$500 and $700 million”. In March 2025, New Brunswick’s Finance and Energy Minister reportedly admitted that the province was not planning to invest any more in ARC-100 without “a clear indication that it’s moving forward fast enough.”1518

The other SMR designer that has received significant financial support from the government of New Brunswick and the federal government is Moltex (see Small Modular Reactors: Canada in WNISR2022). In April 2025, Moltex also admitted to financial problems because its U.K.-based parent company Moltex Energy Ltd. faced “a risk of insolvency”1519 and is currently for sale.1520

Given these problems, it is not surprising that NB Power is unsure that any of these two SMRs, in particular ARC-100, which was more favored for earlier deployment, will be available by even the late 2030s.1521 New Brunswick’s government is now also considering building a large CANDU reactor at Point Lepreau and is seeking federal government support for this effort.1522

China

China is operating or building two kinds of SMRs, a high-temperature gas-cooled reactor design called the HTR-PM and an integral pressurized water reactor design, the ACP100 (or Linglong One).

HTR-PM Design

The HTR-PM is often described in superlative terms. When the reactors were declared as operating commercially in December 2023, roughly 10 years after the first pour of concrete (see HTR-PM Design in WNISR2023), the China National Nuclear Corporation (CNNC) announced that China had “completed the world’s first commercially operational modular nuclear power plant with fourth-generation nuclear technology, marking the transition of fourth-generation nuclear technology from experiments to the commercial market.”1523 But any objective evaluation of the commercial viability of this technology would come to a rather negative conclusion.

To start with, the power output of the reactors was reduced by 25 percent of the design power capacity, from 200 MW to 150 MW. There is no publicly available explanation about why the power output has been reduced. At even this lowered output level, the HTR-PM reactors reportedly have been generating only a fraction of what they could have theoretically produced, with a cumulative load factor of only 26.9 percent according to the IAEA’s PRIS database. However, this calculation is misleading on two counts or simply incorrect. First, the 2023 load factor is given as 100 percent even though the units were online only for the month of December 2023.1524 Second, and more importantly, the load factor has been calculated on the lowered power rating rather than what the owners paid for, namely the design power rating. In summary, the HTR-PM’s performance—as published—does not suggest that the design is reliable or profitable.

It is thus somewhat surprising that the Chinese technology promoters plan to address the challenges of the first HTR-PM through the HTR-PM600 project that would see the expansion of the modular concept to six 100-MW modules driving a single turbine. This demonstration project is planned to supply industrial steam to nearby petrochemical facilities in Lianyungang.1525 See China Focus.

ACP100 Design

The ACP100 integrated Pressurized Water Reactor (PWR), also referred to as Linglong One, has been in the developmental phase since 2010, and its initial design was finalized in 2014.1526 Construction of the 100-MW Linglong One started in July 2021 at the Changjiang site in Hainan province, where two CNP-600 PWRs are operating and two Hualong One units are being built.1527 As detailed in WNISR2022, the start of construction was delayed by at least six years. The planned construction period is 58 months, which would mean that the reactor is to become operational by May 2026.1528 The projected construction period is typical for Chinese nuclear reactors, which does not suggest any advantage resulting from building an SMR.

France

France continues to develop the Nuward design, first revealed in September 2019.1529 The initial design was a plant with two modules of 170 MW capacity each.1530 However, this design was not a high priority, at least in the early years. Indeed, in 2020, Nuclear Intelligence Weekly reported that Électricité de France (EDF) “remains almost exclusively focused on developing a more economic version of the EPR in the EPR2,” including the fact that there were “as many as 1,000 engineers involved” in developing the EPR2 in comparison to “10-15 engineers” on Nuward.1531

In July 2024, EDF announced that Nuward was to become “a design based on proven technology bricks only” with the hope that this “orientation will provide better conditions for success by facilitating technical feasibility and reducing risks.”1532 There were technical problems with the earlier design and its market viability.1533 Reuters traced the design changes to feedback from “prospective clients such as Vattenfall, CEZ and Fortum” who wanted to be confident about cost projections and delivery deadlines so that the “levelised cost of electricity for the SMRs would be in the range of 70 to 100 euros per megawatt-hour [about US$80–115].”1534 Four days after the announcement about the change in the design’s orientation, TechnicAtome, which makes France’s nuclear submarine reactors, decided to stop its participation in the Nuward project.1535

In January 2025, Nuward announced that its new design “will deliver 400 MW of power and offer an option for cogeneration, up to approximately 100 MWth”, and a new Executive President, Julien Garrel, an EDF veteran, would drive this “new phase” working towards finalizing the conceptual design by mid-2026.1536

The Nuclear Energy Agency’s SMR dashboard now describes the Nuward design as an “Integrated pressurised water-cooled SMR providing up to 400 MWe” and generating up to 1000 megawatts of thermal power.1537 Despite these changes and a completely new design being developed, in February 2025, the Nuward President declared in a webinar that the schedule envisioned a “conceptual design” by 2026, a “basic design” by 2029 and from 2030, the company would start working on the “detailed design”.1538 Implementation of a Nuward prototype seems far away.

India

As with other countries, India’s government too has extolled small modular reactors, making claims about building many of them.1539 In March 2025, the government announced in parliament that the Department of Atomic Energy (DAE) was developing three kinds of reactors:

• one termed Bharat Small Modular Reactors (BSMR-200) that is intended to repurpose “retiring thermal power plants & captive power plants for energy intensive industry such as aluminum, steel, metal,”
• one termed Small Modular Reactors (SMR-55) for “providing energy for remote as well as off-grid location(s)” and
• one “High Temperature Gas Cooled Reactor of capacity 5 MWth for hydrogen generation”.1540

Although the government indicated these purpose-oriented locations, it also specified that lead units of these reactors are “planned to be set up at DAE sites” and also promised that these “demonstration reactors are likely to be constructed in 60 to 72 months” once they are funded.1541 This followed the government’s announcement the previous month, during the presentation of the 2025–2026 budget, that there would be “at least five indigenously developed SMRs… operationalized by 2033.”1542

In addition, there have also been announcements about non-indigenous, imported SMRs. During Prime Minister Modi’s visit to France in February 2025, both countries signed a Declaration of Intent on Establishment of Partnership on Advanced Modular Reactors and Small Modular Reactors.1543 Earlier, in March 2024, a Rosatom spokesperson was quoted as saying that the company was “in talks with India for small modular reactors”.1544

Another company that has been seeking a market in India is Holtec. The CEO of the U.S.-based business met with the Indian Prime Minister in November 2024 and the company announced in March 2025 that it had received U.S. Government authorization to sell its SMR-300 design and that “three Indian companies – Larsen & Tubro (Mumbai), Tata Consulting Engineers (Mumbai) and the Company’s own subsidiary, Holtec Asia (Pune)” were eligible to receive technical information.1545

The listing of private companies by Holtec is not coincidental. As discussed in the section on India, the government aims to attract private companies to invest in nuclear power and build SMRs. However, private companies are unwilling to be held liable for potential accidents and would like to be completely indemnified.1546 Holtec, in its announcement about U.S. government authorization, complained about India’s liability law labelling it a “legal barrier”.1547

Whether it is small or large reactors, India’s plans to import technology have mostly not materialized. Nor has it achieved its targets for nuclear energy deployment. These factors, and the many challenges confronting SMRs, in particular their lack of economic viability, render the ambitious plans about building SMRs also quite unrealistic.

Russia

Russia is building both light water and fast neutron SMRs, with a special focus on barge-mounted reactors for coastal locations also referred to as “floating” nuclear reactors or power plants.

Light Water Reactor Designs

Russia operates two KLT-40S SMRs loaded on a barge called the Akademik Lomonosov that were commissioned in May 2020 after lengthy delays and cost overruns that have been detailed in earlier WNISR editions. The two reactors have not been operating very efficiently, with lifetime load factors of just 36 and 28 percent until the end of 2024, as reported by the IAEA-PRIS database.

At least two more SMR projects based on light water reactor designs are underway. The first is in the Chukotka Autonomous region and involves four 55 MW RITM-200S reactors, and the second is in Yakutia involving a (land-based) RITM-200N reactor design.1548 However, as explained in Russia Focus, it is “difficult to determine the exact number of floating nuclear units currently under active construction” because there is no “established terminology and clear milestones for floating nuclear plants.” None of these SMR projects are listed as under construction by the PRIS database.

That said, the start date for the first of the RITM-200S reactors has been announced variously as 2028,1549 and “by 2029”.1550 Keel laying for the barge that is to hold two RITM-200S reactors commenced in August 2022.1551 The barge is being built in China, by Wison (Nantong) Heavy Industries, which in 2021 won the contract at a reported price of US$226 million.1552

As of mid-2024, the Yakutia SMR project was projected for start in 2028,1553 however in October 2024, the target date was revised as 20311554. Back in 2020, Rosatom had projected a commissioning date of 2027 for a domestic land-based RITM.1555 In March 2025, Rosatom was reported to be producing eight RITM series reactors at its ZIO-Podolsk plant, but some of these are for icebreaker vessel propulsion.1556 However, as detailed in Russia Focus, Russian news agency TASS announced two months later that there were ten RITM series reactors, including six RITM-200 and four reactors for icebreakers, being manufactured.1557

Fast Neutron Reactor Design

Russia has long been interested in fast neutron reactor designs and breeder reactors.1558 But, when it comes to SMRs, the only design it is building is the lead-cooled BREST-300, which has been under construction since June 2021 at the Siberian Chemical Combine (SCC) in Seversk.1559 Originally, the reactor was to have started operating in 2020.1560 As of the end of 2024, the anticipated grid-connection date for the reactor was 2028,1561 and the cost of the reactor, according to one source, was RUB100 billion (US$20211.4 billion).1562 There seem to be no plans to build more BREST-300 units and Rosatom is contemplating following up with a 1200 MW unit, i.e., not an SMR.

Export Prospects

Rosatom also continues with its plans to export SMRs. Over the past year, Rosatom has made numerous announcements about specific countries that it envisions exporting SMRs to, including Mongolia, Indonesia, and Kyrgyzstan.1563 However, the only contract it has to build SMRs is with Uzbekistan that was signed in 2024.1564 Initially, the project envisioned constructing six RITM-200 SMR units.1565 However, in June 2025, Rosatom announced that the project had been reconfigured to building two RITM-200 units and two large (gigawatt scale) units.1566

South Korea

According to an October 2024 IAEA report, South Korea now has “two notable SMR designs”, namely the System-integrated Modular Advanced Reactor (SMART) design and the innovative SMR (i-SMR) design.1567

As detailed in earlier WNISR editions, SMART is a 100-MW pressurized water reactor design that was licensed in 2012, but has failed to win any domestic orders because of its high estimated construction cost per unit of capacity. Nothing has come so far either of South Korea’s attempts to export the SMART design.1568

A modified version of the SMART, called SMART100, was then developed. In December 2019, the Korea Atomic Energy Research Institute (KAERI), Korea Hydro & Nuclear Power (KHNP) and Saudi Arabia’s King Abdullah City for Atomic and Renewable Energy (KA-CARE) applied to South Korea’s Nuclear Safety and Security Commission (NSSC) for a standard design approval of the SMART100. The NSSC began its review of the application in August 2021 and granted the approval in September 2024.1569

The other design the “i-SMR”, has become more prominent in recent years. Over the past year, this design has been the focus of government funding. As detailed in South Korea Focus, the government allocated KRW85.9 billion (US$202481 million) towards the i-SMR and unspecified SMRs. Earlier, in July 2023, the Ministries of Trade, Industry and Energy, and Science and ICT [Information and Communication Technologies] announced KRW399 billion (US$2023306 million) in funding to establish the “Innovative [SMR] Technology Development Project” with the goal of obtaining the Standard Design Approval (SDA) by 2028.1570 The i-SMR Development Agency envisions approval by 2028, site preparation to start by 2029 and construction taking place from 2031 to 2034. The “standard design” of the i-SMR is to be completed by December 2025.1571

Even before the design is completed, KHNP signed MoUs with Indonesia’s PLN Nusantara Power and the Jordan Atomic Energy Commission to explore deploying i-SMRs in these countries.1572 In April 2025, at a conference in Amman, the head of KHNP’s Central Research Institute went as far as to claim that the i-SMR “offers the optimal solution for Jordan’s future energy needs.”1573 That seems unlikely, especially since Jordan’s policymakers have, in the past, expressed goals of having economically competitive electricity generation costs and not wanting their country to be a guinea pig for new designs.1574

The other SMR designs that South Korean organizations have pursued are ones aimed at marine propulsion and floating nuclear power plants.1575 None of these have been submitted to the NSSC for safety evaluation or licensing so far. But in February 2025, a Danish company announced that it was “ready to start licensing activities in South Korea.”1576 Saltfoss (then Seaborg), had formed a consortium with KHNP and Samsung Heavy Industries in April 2023 to develop floating nuclear power plants, and as of early 2025 the partners planned for a “first reactor to be ready in the first half of the 2030’s, followed by series production from the mid-2030’s.”1577

United Kingdom

In June 2025, the British government picked Rolls-Royce “as the preferred bidder to partner with Great British Energy–Nuclear to develop small modular reactors, subject to final government approvals and contract signature.”1578 As part of the announcement, the U.K. Government also pledged “over £2.5 billion (US$3.4 billion) for the overall small modular reactor programme”. The funding for this will come out of the £8.3 billion (US$11.2 billion) budget of Great British Energy that was originally viewed as a vehicle for investing in renewable energy.1579 Earlier in the year, Siemens announced that it would be “the sole supplier of steam turbines, generators, and other auxiliary systems” for the SMRs to be built by Rolls-Royce.1580 A final contract is to be concluded by the end of 2025.

The selection process was announced in July 2023 by then Energy Security Secretary, who called upon companies to “register their interest with GBN [Great British Nuclear] to participate in a competition to secure funding support to develop their products” which could “result in billions of pounds of public and private sector investment in small modular reactor (SMR) projects in the UK.”1581 The process was meant “to identify the best, most appropriate, SMR technologies” and the financial support was to help companies whose designs are selected to further develop their technologies in preparation for “a Final Investment Decision” to be made by 2029.1582

The Rolls-Royce SMR design dates back to 2017 and it was initially rated at 440 MW of electricity,1583 i.e., not really meeting the definition of a small reactor as one designed to generate under 300 MW of power. By 2021 the Rolls-Royce SMR design was further uprated to 470 MW, and its then-Chief Technical Officer traced the increase to a desire “to minimise the cost of energy coming out…the cost being the historical challenge of nuclear power.”1584

Even before the government’s decision to choose Rolls-Royce, Westinghouse, one of four companies still in the running at the time, reportedly withdrew from the competition.1585 Westinghouse’s decision did prompt surprise because of its announcement, from only a few months earlier, in September 2024, that its design had been “downselected for the final phase” of the British competitive SMR technology selection process, claiming that this amounted to a recognition of “the advantages of the AP300 SMR design” that ensure that the “AP300 can get to market quickly, economically and with certainty.”1586 However, one analyst has calculated that the U.K. Government’s financial commitment was insufficient to allow for “fleet mode” factory production of SMRs, and this might be a factor in Westinghouse’s decision.1587 Earlier, in July 2024, EDF withdrew its Nuward design from the competition.1588

Last year, in April 2024, the Office for Nuclear Regulation (ONR) was “currently carrying out Generic Design Assessments (GDAs) on Small Modular Reactor (SMR) technologies proposed by Rolls-Royce SMR Ltd, Holtec International and GE-Hitachi Nuclear Energy International LLC.”1589 ONR announced in July 2024 that the Rolls-Royce design had completed the second step of the GDA process and that it could “now continue into Step 3 assessing in more detail the evidence that supports the claims made about the design in the Step 2 submissions.”1590 The next month, ONR announced that Holtec’s SMR design had completed the first step of the GDA process, and that it would “begin the technical assessments of the reactor” as part of Step 2.1591 A nearly identical announcement about GE-Hitachi’s BWRX-300 came in December 2024 and ONR explained that during Step 2 “regulatory activity will be targeted on assessing the fundamental adequacy of the GE-Hitachi BWRX-300 design for deployment in Great Britain” and this will “consider the suitability of the methodologies, approaches, codes, standards and philosophies identified by GE-Hitachi in the generic safety, security, safeguards and environment cases for securing future regulatory permissions and permits.”1592

United States

The U.S. Government has been the single largest funder of SMRs, in particular on research and development. Over the past year, the U.S. Department of Energy (DOE) was the largest source of funding. As detailed in United States Focus, a combined US$2.8 billion in matching grants has been awarded to X-energy, TerraPower, and Kairos Power through the DOE’s Advanced Reactor Demonstration Program. In March 2025, the DOE reissued a US$900 million solicitation to support the deployment of light-water SMRs, with US$800 million earmarked for “up to two first-mover teams of utility, reactor vendor, constructor, and end-users or power off-takers committed to deploying a first plant while at the same time facilitating a multi-reactor, Gen III+ [Generation III+] SMR orderbook” and up to US$100 million to fix “key gaps” troubling the domestic nuclear industry in areas including “design, licensing, supplier development and site preparation”, with the goal of spurring additional Gen III+ SMR deployments.1593 The funding was first announced in June 2024 by the earlier Biden administration.1594

Based in part on this large amount of public funding, there is growing interest by other stakeholders in building SMRs. In May 2025, the Tennessee Valley Authority (TVA) submitted a permit application to the U.S. Nuclear Regulatory Commission (NRC) to build a GE Hitachi BWRX-300 SMR design at the Clinch River site near Oak Ridge, Tennessee.1595 TVA is also part of a coalition applying to the DOE for US$800 million “to accelerate BWRX-300 development” as well as US$8 million to support the NRC license review cost. This is not TVA’s first attempt to plan for SMRs. Back in 2011, it planned to construct the leading SMR model of that time, the mPower SMR design, at the Clinch River site and two years later it signed a contract with Babcock & Wilcox, the company developing the mPower design, to prepare and submit a construction permit application to the NRC.1596 However, the project was terminated in 2017.1597

Also, in the state of Tennessee, Kairos Power started the first pour of concrete in May 2025 for the Hermes low-power demonstration reactor.1598 Hermes, which will not generate any electricity, received the permit for construction from the NRC in December 2023.1599 However, Kairos Power has also received a permit in November 2024 to build a facility with two 35-megawatt thermal test reactors called Hermes-2 with a shared power conversion unit.1600

In May 2025, the NRC issued a Standard Design Approval for the 77-MW NuScale design US-460.1601 This design represents a more than 50 percent increase in power compared to the 50-MW NuScale design that received the more stringent Standard Design Certification from the NRC in 2023.1602

Earlier, in March 2025, Long Mott Energy LLC, a subsidiary of Dow—which now describes itself as a “Materials Science company” rather than a Chemical company—and X-Energy applied for a construction permit to the NRC to build four 80-MWe X-energy SMR units in Seadrift, Texas.1603 In 2022, when Dow entered into an agreement with X-energy, the announcement said that the X-100 reactors were “expected to be operational by approximately 2030,”1604 whereas the filed application reads “the completion date for construction is expected to be no later than 2033.”1605

Another SMR design that is being evaluated by the NRC is TerraPower, which submitted a construction permit application in 2024.1606 The NRC listed around 50 items that needed to be addressed.1607 In the meanwhile, as described in United States Focus, the developer broke ground on a “test and fill facility” for the reactor’s sodium coolant/moderator in June 2024.

Media coverage about SMRs in the United States has largely focused on their potential use for powering data centers used for computations related to Artificial Intelligence (AI). This has been fueled by investment and partnership announcements by cloud companies, in particular Amazon, Google, Microsoft, and Meta.1608 However, the levels of announced investments are small and completely inadequate with relation to how much is needed to build a nuclear reactor.1609

Conclusion

Despite much hype, Small Modular Reactors are not living up to the ensuing expectations. The evidence from the cost estimates of SMR projects produced so far show that they are not to be expected to produce economically competitive electricity. One authoritative estimate of the cost of electricity from SMRs comes from Australia’s Commonwealth Scientific and Industrial Research Organisation (CSIRO) showing that SMRs would be by far the most expensive way to generate power in the country; for example, the Levelized Cost of Energy (LCOE) for an SMR project starting to deliver power in 2030 was estimated at AUD328–619/MWh (US$211–399/MWh); in comparison, solar PV and onshore wind were estimated at AUD35–63/MWh (US$22.5–41/MWh) and AUD 64–107/MWh (US$41–69/MWh) respectively.1610 When integration costs are included to account for the variability of wind and solar power, the cost of electricity in 2030 was estimated to be in the range of AUD90–131/MWh (US$58–84/MWh) even when variable resources constitute high levels of the electricity supply. According to the energy economist who led CSIRO’s analysis, the capital costs for SMRs used by CSIRO closely matches the estimates for the BWRX-300 reactors to be built at the Darlington site in Canada.1611

SMR advocates often look to factory production to make reactors cheaper in their models. For example, the IAEA claims that SMRs “show the prospect of significant cost reduction through modularization and factory construction which should further improve the construction schedule and reduce costs.”1612 Likewise, a NuScale official told a conference on SMRs in 2016 that the company’s vision was “to see NuScale Power Modules rolling off production lines in British factories and generating power for British homes in the 2020s”.1613

In reality, the production lines are likely to be much slower than can be described as “rolling off”. When Holtec announced in May 2025 that they had selected a site for their U.K. factory, the press release revealed their proposed factory would produce “2 SMR units annually”,1614 which sounds a lot more like crawling off than rolling off. At this slow rate of production, it is unlikely that there will be significant cost savings because reductions in cost through learning depends in large part on numbers of units produced.

Challenges of Integrating Nuclear Power into the Energy System

Introduction

For years, proponents of nuclear energy have been pushing for a revival, claiming it is both imminent and necessary. The validity of this assertion can be verified by observing the development of the nuclear industry, technological advancements, and relevant investments. The WNISR has been accomplishing this task for many years by delivering detailed ex post or real-time analysis. In order to better understand the underlying reasons for current and likely future trends, this chapter undertakes a more fundamental comparison of the characteristics of nuclear technology and its competitors.

Setting the Stage

Assessments of future developments depend on the understanding of their historical context. Economic and technological developments follow their own trajectories, so it is essential to consider the historical evolution of this context. The next section of this chapter clearly shows that the context for civil nuclear power has indeed changed dramatically in the seventy years since its first commercial deployments.

Reframing the Question

After an early period of rapid, policy-driven growth, for decades, nuclear power deployment and economics have stagnated or declined in most countries. Nuclear power is facing new competition from new renewable technologies, especially photovoltaics and batteries, supported by power electronics, which are experiencing relentless cost reductions and massive deployment. Is this difference due to fundamental factors and will it persist, or is a revival of nuclear power at least conceivable in principle, perhaps depending on a boost from political and financial support?

The Approach Taken

This chapter endeavors to answer this question by means of two comparisons.

In a first step, nuclear and renewable technologies are compared in terms of their fundamental technological and physical characteristics and the systemic implications of their characteristics. This comparison will provide an overview of the actual energy supply units and their supporting systems in terms of their characteristics, costs, and development prospects.

The second step involves comparing the roles of these individual energy production systems within the overall energy supply and consumption system. This will provide insight into the ability of the two fundamentally different types of technology to meet the needs of the emerging overall energy system and their compatibility with each other—i.e., the extent to which they can coexist.

This chapter does not compare nuclear power with end-use efficiency—another formidable competitor.

Key Findings

A thorough examination reveals that nuclear energy exhibits notable drawbacks when contrasted with renewable energy technologies across both comparative levels. Examining the underlying physics of the two technologies shows that significant disparities between them render nuclear inherently more costly, both in its present state and in the future. An analysis of the system characteristics of the emerging overall energy system indicates that photovoltaics, batteries, and power electronics are already modifying the logic of the overall system in a manner that renders it increasingly challenging for nuclear power to be integrated into. Time horizons play a key role: until a nuclear power plant, decided upon today, will start producing electricity, new technologies will have thoroughly transformed the energy system.

Radical Changes in the Context of the Commercial Use of Nuclear Power

The technological, economic, and institutional foundations of the global energy system have undergone profound changes in recent decades, especially in the past ten to fifteen years. These changes form a radically different context from that in which the nuclear industry first emerged and developed. Understanding this changing landscape is essential for any comparative assessment of energy technologies and their respective roles in future energy systems.

The Context Sixty Years Ago

The conceptual and institutional architecture of nuclear energy was largely formed during the 1950s, 60s, and 70s. This period was marked by expectations of ever-increasing energy demand and a strong belief in the ability of large-scale engineering projects to deliver progress.

In the formative decades of nuclear power, electricity generation was dominated by large power plants: coal- and oil-fueled thermal power stations as well as hydroelectric power generation. Both technologies strongly favored large units due to their intrinsic physical properties. Coal prices were rising. Oil reserves seemed limited and were concentrated in a few countries. The discovery and use of natural gas was growing. A new power source was welcome.

Electricity supply was organized by strong monopolies, in many countries under state control or state regulation. Power generation costs were not transparent to the public. Grid stability was ensured by adapting the generation in large units to variable demand and by significant safety margins in operating essentially untransparent grids.

The commercial use of nuclear power was originally developed and promoted by state agencies as a further and more publicly acceptable use of military nuclear technology. Large subsidies and liability exemptions were necessary to convince private companies to engage in the development and deployment of civilian nuclear technologies.1615 However, the development of military and civilian uses of nuclear energy remained strongly interlinked.

While nuclear power was not cost-effective, it was widely hoped that costs would come down with learning and that safety challenges could be managed. A broad range of reactor concepts was developed in different countries. However, since the 1960s, there has been growing criticism against nuclear power—first mainly motivated by the link to military technologies, then increasingly by safety issues and concerns over radioactive waste. Public debate and increasing insight caused research and state institutions to develop major research programs into nuclear safety and detailed licensing procedures contributing to increasing costs.

Increasing Climate Awareness and the Urge for Decarbonization

Already in the 1970s, the problem of human-made climate change was known and used as an argument for developing nuclear power. However, even in the heydays of nuclear power deployment, it could not live up to this challenge: By 1997, when the UN Framework Convention on Climate Change was finally set up with the signature of the Kyoto Protocol, the construction of nuclear power plants had already declined to its late 1950s levels (see Figure 8). The share of nuclear power in primary energy consumption then reached 6.5 percent. In 2023, 26 years later, primary energy consumption had risen by 64 percent, and nuclear power had not contributed to the meagre reduction of the share of fossil fuels (from 86 to 82 percent): It was instead solar and wind that rose from zero to 6 percent, while nuclear declined from 6.5 to 4 percent with a stagnating absolute input.1616

In the past two decades, the climate problem has climbed to the top of the energy agenda with “energy security” concerns increasingly driving policy decisions. Most countries and industries are considering decarbonization as an urgent necessity. After having suggested nuclear power as a solution to the oil crisis in the 1970s, to air pollution in the 1980s, to alleviate global poverty in the 1990s, and to fuel scarcity in the 2000s, nuclear proponents are proposing their technology as a key solution to decarbonization, despite its poor track record in combating climate change and its ongoing cost increases. Indeed, the carbon footprint of nuclear power is much lower than that of fossil fuels. But do advocates of nuclear energy have new arguments vis-à-vis the upcoming clean alternatives? The key questions are whether nuclear power is more cost-effective in reducing emissions and how fast it can be implemented.

Step-by-Step Liberalization of Electricity Markets

In the 1970s, the upcoming use of combined heat and power (CHP) generation in industry started to undermine the power of monopolies. While large thermal power plants lose half to two thirds of the input energy as waste heat that needs to be cooled away, new small gas turbines offered cheaper electricity by simultaneously delivering heat for industrial purposes. In 1978, new U.S. energy legislation permitted independent electricity producers to sell electricity to the existing monopolies. This led to the emergence of a growing number of independent power producers, who utilized not only combined heat and power or cogeneration but also early versions of wind power. Increasingly, prices were being decided using market-based tools like auctions. This development was not unique to the United States. Denmark and the Netherlands are outstanding examples of an early and intensive adoption of cogeneration schemes.1617

While in the U.S. electricity system, the growing fragmentation of political and economic power led to inconsistent and divergent approaches across the states, the European Union embarked on a much more ambitious liberalization. As in the U.S., the emergence of efficient distributed generation technologies, combined with a growing environmental movement on the one hand, and the increasing influence of neoliberalism on the other, led to an increasingly untenable position for the old electricity monopolies. Meanwhile, Western European electricity markets had become increasingly interconnected, and with the opening of the Iron Curtain in 1989, the integration of Central and Eastern European countries into the European single market was on the horizon. From the late 1990s, this led to the gradual liberalization of European electricity markets and the introduction of national and European regulators, thereby creating the largest liberalized electricity market worldwide. However, the implemented market mechanisms were still designed with the old technologies and control logics in mind.1618

In recent years, it has become increasingly clear that shaping liberalized markets in an environment where natural monopolies (electricity grids) and security of supply play a central role requires sophisticated market design and highly competent oversight. An additional objective has gained importance with the increasing push for decarbonization. Moreover, the very different characteristics of new energy technologies require a recalibration and redesign of markets designed for a traditional technological environment. As a result, the role of regulators and an intense political debate on the objectives and modalities of the energy transition have become more important.

New Energy Technologies Start to Disrupt Markets and Systems

The chapter Nuclear Power vs. Renewable Energy Deployment illustrates the unprecedented momentum with which solar power, wind energy, and batteries are being deployed and disrupting markets. The declines in unsubsidized costs observed in photovoltaics, batteries, and power electronics over the past decade are unparalleled in the history of energy technologies. Costs have declined for decades, but only recently have competitiveness thresholds been crossed, leading to disruptive transformations in the energy economy. Figure 59 in that chapter is a snapshot of how fast solar power generation is outpacing nuclear power and shows the remarkable complementarity between solar and wind. The reason for the new technologies’ capability to achieve such progress compared to fossil fuels and nuclear power is rooted in fundamental physical properties, which will be explored in the next section.

Moreover, the new energy technologies, unlike nuclear power, are not simply substitutes for the old, well-known fossil power plants. They are fundamentally changing the logic of the energy system. The combination of unprecedented characteristics of photovoltaics, batteries, and power electronics leads to a transformation of the whole energy supply system, involving not only generation but all steps until the end use:

• Electricity generation becomes more and more decentralized, involving millions of investors and operators. Generating technologies shift from giant and bespoke to modular and mass-produced.
• Electricity becomes the central, universal energy carrier in the overall energy system.
• Electricity generation with sunlight and wind depends on varying weather, daytime, and seasons, requiring the use of new flexibility and storage options across the entire energy system.
• Increasing difficulties in controlling this complex system with centralized approaches give way to new bottom-up control logics.

The section after next will further explore these changes in the systemic context brought about by new energy technologies, and how they affect the prospects of nuclear power.

Widening International Competition

The liberalization of electricity markets and international trade for equipment, in combination with new technological developments, opened opportunities for new actors in the electricity and technical equipment markets.

• New competitors emerged in the national electricity markets: independent power producers, traders, providers of special services to utilities, and large consumers (today, European energy market regulations distinguish 49 different market roles1619).
• Neighboring industries using innovative technologies entered the equipment and services markets: producers of cogeneration engines, wind turbines, digital controls and corresponding software, energy-efficient consumer products, and home energy management systems.
• Competitors from Asian countries, in particular China, started to conquer equipment markets with new energy technologies. Today, China controls over 75 percent of the world’s production of solar panels, batteries, and LEDs, as well as important parts of global markets for power electronics and control device components.1620

The entrance of new actors in the energy technology markets has upended historical oligopolies of European and American companies and has accelerated competitive innovation.

Not only has technological and market competition between companies and industrial sectors increased, but also regulatory competition between countries. Energy policy has become a key element in the competition between political and economic blocs. However, business competition and regulatory competition operate on different time scales. Large public infrastructures have a long-lasting influence. Creating long-lasting dependencies can also be a geopolitical objective—as can be observed in Russian export efforts for nuclear power plants.

Accelerating Change

The number of construction starts of nuclear power plants peaked in 1976. Consequently, the installed nuclear power capacity exhibited the most substantial growth between 1980 and 1985, attaining an average annual growth rate of 13 percent.1621 At a time when nuclear energy competed with coal power plants, this was an exceptional occurrence. Meanwhile, change has accelerated: Between 2019 and 2024, the installed capacity of photovoltaics reached an average annual growth rate of 28 percent.1622

From 2015 to 2024, the average construction time of a nuclear power plant—from first concrete for the reactor-building base-slab to grid connection—was 9.4 years, representing an increase from less than eight years in the 1980s. In contrast, it takes approximately two years to commission a PV plant—the actual construction time may often be only a few months.1623

The International Energy Agency’s (IEA) Net Zero Emissions scenario, which aims to fully decarbonize the energy system, requires non-fossil power generation to grow by an average of 14.5 percent between 2023 and 2030.1624 In such a context, increasing or even maintaining its electricity market share is out of reach for a nuclear industry that, over the past decade, has struggled to sustain its generating capacity and on average takes almost a decade to build a new plant—much more if planning, licensing, public inquiries, and site preparation are included. Even the IEA’s assumption that nuclear capacity could grow by an average of 4.2 percent per year from 2023 to 2030 (still resulting in a 48 percent decline in its share of non-fossil electricity) seems far from realistic.

As a result of all these changes, decision-making in the electricity and the broader energy sector has considerably evolved. Planning horizons have shrunk, decisions need to be taken faster, the regulatory environment is changing rapidly, and old oligopolies have lost control.

In the aftermath of Russia’s full-scale invasion of Ukraine, there has been a shift in energy policy goals: Energy supply security has gained weight over economic and climate change concerns. System resilience, the avoidance of dependencies, and the role of large infrastructures are seen in a new light. This means that both longer-term considerations and higher flexibility are getting more attention.

Physical Principles Underlying the Technologies

Nuclear Energy

Two Fundamental Dangers of Exploiting the Bonding Forces in Atomic Nuclei

Nuclear energy is the only energy generation technology that deals with the bonding forces in atomic nuclei, mainly the so-called strong interaction, one of the four elementary forces in physics. This fact has far-reaching consequences.

All other technical energy sources rely on harnessing the forces of gravity and electricity. Wind energy and hydro collect mechanical energy of natural phenomena at the macroscopic scale that originate in the transformation of solar electromagnetic radiation into heat. In combustion, chemical reactions change the bonds between atoms provided by the electric forces in the electron shells of atoms and molecules, and then deliver heat (random movement of molecules), which in further steps may be converted into mechanical and electric energy. In photovoltaics, the input of sunlight (electromagnetic radiation) liberates electrical charge carriers in the electron shells, which, if appropriately collected, directly provide electrical current.

Some Essential Nuclear Physics

Nuclear energy, indeed, touches forces that have never been dealt with before. In 1911, Ernest Rutherford concluded from a series of experiments that the matter in an atom is not evenly distributed. Rather, it is concentrated in an extremely small, positively charged nucleus, which is surrounded by a much larger shell of orbiting, nearly massless, negatively charged electrons. Step by step, it became clear that the nucleus was essentially composed of positively charged “protons” and electrically neutral “neutrons” (discovered in 1932), and that the electric repulsion among the protons must be overcome by a new, until then unknown, extremely strong cohesion force with much shorter reach than electrical forces.

This early model of nuclear physics, supported by a series of revolutionary scientific discoveries of that time and the evolving mathematical formulation of quantum physics, enabled the explanation of the perplexing phenomenon of radioactivity—a new kind of radiation discovered in 1896. It occurs when unstable atomic nuclei decay, emitting alpha particles (two protons and two neutrons), beta-radiation (electrons), gamma-radiation (hard electromagnetic waves), neutrons, or combinations of these. The timeline of radioactive decay processes follows stochastic laws: every kind of nucleus, or isotope (characterized by the number of protons and neutrons it contains), has a specific, constant half-life (the amount of time after which half of the original material has decayed) ranging from microseconds to millions of years. In particular, this radioactive decay can be observed in all nuclei heavier than lead, which are unstable because of their high proportion of neutrons. They may even split into several larger parts (nuclear fission), releasing kinetic energy of the fission products and emitting neutrons. These neutrons are being absorbed by other nuclei, which consequently may become unstable themselves.

As early as 1933, the Hungarian physicist Leó Szilárd, who had emigrated from Berlin to London, conceived the idea of a self-sustaining nuclear chain reaction that could unleash vast quantities of energy from relatively small amounts of fissile material. In 1936, he filed a patent for this concept, which was kept secret due to its military significance. Later, joined by the more famous Albert Einstein, he warned U.S. President Roosevelt that Germany could build a nuclear bomb. This motivated the huge effort of the Manhattan Project. It was only in the aftermath of the nuclear bombings of Hiroshima and Nagasaki—when the global community realized the potential of self-destruction and both sides in the Cold War were heavily engaged in nuclear armament—that the idea of a civilian use of nuclear power was intensely propagated.

The chain reaction used in most nuclear power reactors today relies on the fission of Uranium-235. Upon absorbing a neutron, it transforms into Uranium-236 and splits into a range of radioactive fission products, emitting three neutrons. Such a self-sustaining chain reaction was first realized by Enrico Fermi in 1942 under strict secrecy, accomplishing a key milestone in the Manhattan Project.

Dealing with Unavoidable Long-Term Radiation

Exploiting nuclear forces by splitting and thereby changing the configuration of atomic nuclei inevitably involves the release of intense radiation. More importantly, the fission products, whose kinetic energy makes up for most of the energy released, are radioactive themselves. Their half-lives extend over an extremely large range, from tiny fractions of a second up to millions of years. Moreover, the absorption of excess neutrons by neighboring nuclei results in the transformation of stable material into unstable isotopes. This phenomenon increases the quantity of radioactive material. As a result, the use of nuclear power is inevitably accompanied by

• radiation of different kinds, which, depending on its intensity, can be extremely harmful to living beings;
• radioactive material in solid, liquid, and gaseous forms, which, depending on the isotopes involved, may radiate for up to millions of years.

Depending on the type of radiation (alpha, beta, gamma, neutron), quite different barriers are needed to shield against its effects. In some cases, it is even more challenging to prevent the inhalation or ingestion of small amounts of radioactive material that may be incorporated into living organisms and radiate there directly for many years or until the end of life of the said organism. However, far more difficult than shielding or containing radioactive materials in current operations is ensuring their safe isolation from living beings for many millennia.

Radiating isotopes cannot be destroyed by chemical processes, such as burning, as these do not alter the atomic nuclei. The decades-old discussion of concepts to transmute long-lasting radioactive isotopes into ones with shorter half-lives, for example, by bombarding them with neutrons, has not led to encouraging results: Just reducing the necessary high-safety storage time for a selection of particularly long-lived isotopes—still to several hundred years—would require considerable efforts involving risky separation processes, complex plants with high energy consumption, many decades of additional reactor operation, as well as additional lower-level radioactive waste.1625 In most cases, it would also lead to increased proliferation risks. According to a major 1996 U.S. National Academy of Sciences study, coordinated by nuclear engineering icons Norman Rasmussen and Thomas Pigford, in the real world, it would take centuries to obtain the theoretical results of transmutation, notwithstanding the fact that a significant share of long-lived wastes is not in a physical or chemical form apt for enhanced reprocessing, separation, and transmutation.1626

The inevitable radioactivity involved in nuclear power generation results in a series of burdensome and costly challenges:

• Operating a nuclear power plant requires adequate shielding from radioactive radiation around all components that are in direct or indirect contact with radioactivity.
• Maintaining a nuclear power plant necessitates specialized equipment, methodologies, and shielding measures, as construction elements exposed to neutron radiation become radioactive themselves. Thus, even simple repairs that would require little effort in a conventional power plant become cumbersome, time-consuming, and costly.
• Not only spent fuel containing fission products, but all materials exposed to radiation or radioactive material must be tightly contained and controlled; this requires special logistics, cooling, and surveillance.
• Depending on their category, nuclear wastes must be safely stored for up to a million years. This requires not only safe containment over long periods, but also protection from malicious attacks. Fissile or radiating material at all stages is a potentially attractive target for military and criminal ends.

These specific requirements, which have become increasingly stringent with the advancement of knowledge, contribute to the construction and operation expenses of private nuclear facilities. However, the temporal extent of the risks associated with nuclear waste far exceeds the temporal scope of liability legislation applicable to other industrial activities. The inevitable result was the assumption of long-term civil responsibility by nation-states (see Nuclear Economics and Finance in WNISR2023).

Controlling a Potentially Explosive Nuclear Chain Reaction

Exploiting nuclear power for civilian purposes relies on maintaining a self-sustaining, stable nuclear chain reaction that produces heat at a constant rate. Appropriate control can be exercised by regulating the absorption of neutrons by nuclei other than the fissile fuel. In standard reactors, this is realized by inserting control rods of neutron-absorbing material into the reactor core. In large reactors, the process must be meticulously regulated in all sections of the reactor core to avert local overheating, a phenomenon that could result in the melting of structural components and loss of control. In the case of local malfunctions, ensuring sufficient cooling is essential for avoiding further damage.

Over the years, reactor safety studies have identified many actual and potential concatenations of incidents. In particular, the severe nuclear accidents at Three Mile Island (U.S., 1979), Chornobyl (USSR, 1986), and Fukushima (Japan, 2011) have spurred reactor safety research. Subsequently, redundant systems and multiple safety layers have been added to the design of nuclear power stations, raising complexity and costs. This also encompasses the periodic and meticulous inspection of all critical components, along with their timely replacement in the event of signs of potential failure, such as the presence of microscopic cracks. Such inspections and replacements increase nuclear plants’ downtime compared to conventional power plants and compete with the economic incentive of steady operation of the plant. Steep load changes may cause wear on the non-nuclear parts and thermal strain on critical components. The power output of nuclear power plants, therefore, is less flexible than that of conventional plants.

As nuclear bomb explosions and nuclear reactor accidents have demonstrated, potential risks in the case of accidents are not only given by the amount of energy released and the immediate destructive effects of radiation, but mainly by the dispersion of radioactive material with long-term effects. Risk is a product of the probability of an event and the potential damage. Even if probabilities of heavy accidents may have been lowered by adding multiple control, emergency, and protection systems, the magnitude of their destruction potential remains such that usual liability and insurance mechanisms are incapable of coping with them. Moreover, there are doubts whether risks have effectively been reduced as increasing complexity and new kinds of dangers (for example, through cyber or drone attacks) have added new challenges. All countries seeking nuclear power have had to introduce exemptions from normal liability rules for industrial operations, thereby socializing risks and reducing operators’ incentives to mitigate them.

A Key Reason for the Rising Costs of Nuclear Power

These interlinked challenges of radioactive radiation and nuclear chain reaction are unique to power generation in nuclear reactors and are intrinsically linked to the manipulation of strong forces within atomic nuclei. With increasing experience and knowledge, risk awareness and the complexity of protective measures have increased. Repeated attempts to reduce complexity with new designs have only had very limited success, as the basic problems are rooted in fundamental physical laws.

Nuclear Power Plants Need to Be Large

Nuclear reactors are built to produce heat. In a subsequent step, the heat is partially converted into mechanical energy using turbines and then into electricity using generators. These turbines and generators are of the same kind as those developed for conventional coal power plants. Temperatures produced by usual nuclear reactors, however, are lower than those reached by burning fossil fuels. The conversion efficiency of heat into electricity, therefore, amounts only to 35 percent or less compared to 46 percent for advanced coal plants and up to 60 percent for gas-fired turbines, whose waste heat then drives steam turbines in “combined-cycle” plants.

Thermomechanical Engines Are Intrinsically More Efficient at a Larger Scale

Due to several physical laws, thermal and mechanical devices are intrinsically more efficient when they are larger. Larger thermal installations have relatively smaller heat losses as volume (heat content) grows more strongly than the surface (losses)—when doubling a sphere’s diameter, the heat content increases eightfold, while the losses rise only fourfold. Moreover, larger hydraulic systems have smaller friction losses: doubling the diameter of an endless pipe reduces flow resistance to 1/16 (or 6.25 percent).1627 In real pipe systems, the effect is even larger, resulting in a reduction by over 96 percent. Moreover, larger turbines allow for higher steam pressures since larger components (e.g., turbine blades) are more robust. The consequence of these and more effects is that steam turbine efficiencies, which essentially reached their maximum in the 1970s, are highly dependent on size.1628

For these reasons, the usual size of conventional coal power plants has reached around 800 MW. Building larger units than these would have too large operational disadvantages in the grid, since substituting the regular output when the plant is stopped for maintenance or other reasons gets more difficult with size. As described above, smaller units can be economical, if not only the power but also the heat output is used (combined heat and power). This, however, presupposes the presence of district heating systems or industrial heat clients in the vicinity. Due to safety concerns, nuclear power plants have invariably been constructed at such a distance from densely populated areas that the economical transportation of heat has not been feasible (with a very few exceptions).

Specific Nuclear Challenges Bring Additional Economies of Scale

Nuclear reactors have additional costs that do not directly depend on the power rating, which is roughly proportional to the volume of the reactor core. Radiation protection, safety containment, etc., depend more on surface than on volume. The costs of operation control systems, access control, or security depend even less on the power rating of the plant. The push for size in nuclear plants is therefore stronger than in conventional thermal power plants. Per kilowatt-hour produced, both operational and capital costs are lower for larger units. The latest generation of European reactors (European Pressurized Water Reactor or EPR) even reaches 1600 MW or more. The two most recent Korean plants are rated at 1340 MW, while Chinese ones are at an average of around 1200 MW.

In addition to being less efficient than fuel-based power plants, this means that nuclear plants must be able to handle unprecedented heat throughputs: the turbine of an EPR has to digest about 2.5 times as much heat energy as the turbines of a large, advanced coal plant.1629

The idea of “Small Modular Reactors”, although not new, has gained significant attention in recent years, with many different design ideas being suggested. The proponents of these concepts claim that smaller, less complex plants with more inherent safety and some degree of serial production can drive down costs. The reality, however, does not yet confirm these hopes. The OECD’s Nuclear Energy Agency has identified 127 SMR technologies.1630 While most of these ideas attempt to revive old concepts from the 1960s and 1970s that have already been extensively discussed and abandoned, the size of the more realistic designs has been increased, and some now approach power ratings of 400 MW—the size of the first serious nuclear power plants in the 1960s. Estimated costs for SMRs keep increasing (for details, see chapter on Small Modular Reactors).

Until now, there is no initiative, the proposed simplification of which has passed detailed planning and all regulatory safety checks without cost increases, which makes a compensation of the physical disadvantages of size reduction illusionary. Even if a new SMR concept were to successfully deal with the need for containing radioactivity and the need for preventing a destructive chain reaction more efficiently than traditional large nuclear power plants, its cost would not be able to compete with photovoltaic power plants which do not have these problems at all.

The increasing amount of critical research on SMRs points to several areas of concern compared to traditional nuclear power plants:1631

• Safety gains are uncertain. Proposed simplifications often lead to unexplored problems or well-known trade-offs.
• Dispersion of the reduced inventory of smaller reactors would still lead to unacceptable destruction—a substantial lowering of the safety level is therefore no option.
• Proliferation and terrorism risks grow with more sites. These risks further increase with higher uranium enrichment levels and the need for spent-fuel reprocessing in several SMR concepts.
• The amount of radioactive waste increases. Higher neutron leakage due to a less favorable surface to volume ratio of smaller reactor cores leads to more structural components being activated.
• Attempts to increase the exploitation of nuclear fuel increase risks and costs in separation processes. Attempts to raise energy efficiency by raising temperatures require the use of other coolants than water, involving new risks. Enabling a more flexible operation by reducing the energy density of the reactor core increases the size of the facility.

Also, selling the waste heat from small reactors (like in small fossil combined heat and power plants) will not make up for the economic penalties of making them smaller.1632 This is not to mention the risks and difficulties of operating such devices in densely populated areas that could use this heat.

Production Series Will Remain Small

Due to the size of nuclear power plants and the complexity of their safety systems, the possibilities of standardization are limited. Specific local conditions—such as options to dispose of large amounts of waste heat through effective cooling systems, the risk of earthquakes and flooding, logistical restrictions, specific security requirements, or the availability of skilled labor—necessitate that nuclear projects be adapted to the specific site and largely single-engineered. Consequently, industrial prefabrication in series is constrained to smaller components that can be standardized.

Chinese, French, and Korean manufacturers have effectively produced small series of nearly identical plants, resulting in cost reductions. These savings stemmed only from a more rational use of known equipment, not from a learning curve with improving technologies. However, this is still far from high-volume industrial production. Even if hundreds of nuclear power plants were to be completed each year, they would still constitute a very small series in terms of industrial mass production. For comparison, a modern solar factory can produce two billion photovoltaic cells per year. In critical areas of a nuclear power plant, while conventional technologies may be used for many parts, the need for specific safety, durability, and quality control requirements necessitates bespoke production from a select number of trusted, specialized vendors. Better cost structures may only apply where off-the-shelf components can be integrated in non-critical parts of the plant.

No Real Innovation on the Horizon

Harnessing revolutionary insights in physics for exploiting the fascinating and extremely powerful forces of atomic nuclei has proven to be loaded with a series of challenges due to fundamental physical principles. The above review shows that nuclear technology has essentially gotten stuck with the concepts of 75 years ago. While the quantum physics revolution has allowed progress in other fields, such as innovative materials, semiconductors, information technology, or artificial intelligence, it has only been marginally useful in advancing nuclear technologies. The share of conventional technology, where little progress is to be expected, remains overwhelming: conventional thermal power generation equipment, physical containment and shielding made of steel and concrete (even if reduced in “advanced concepts”), and cooling systems with pipes, tanks, heat exchangers and pumps make up for most of a nuclear plant’s cost. Recently so-called “new concepts” have mostly turned out to be based on older ideas that have already been discarded.1633 Approaches that could lead to substantial cost reductions are not in sight.

Photovoltaics

The substantial cost reductions in photovoltaic (PV) devices and the electricity they produce have been astounding to most external observers and those who have not closely followed the extraordinary advancements in microelectronics since their invention. The real price of solar modules declined by 99.6 percent between 1976 and 2019.1634 It was William B. Shockley, the 1947 co-inventor of the transistor, who succeeded in 1950 to theoretically explain the photovoltaic effect observed some years earlier in his laboratory when irradiating the boundary layer between differently doped layers of silicon. This quantum-mechanics-based explanation of a potential new energy source, simply using sunlight and semiconductors, motivated further research, resulting in the first two-square-centimeter solar cell in 1954—the same year the first nuclear power plant was connected to a power grid in Obninsk, in the then-Soviet Union.

An Essentially Electronic Technology

Against the backdrop of emerging semiconductor technology, solar energy’s basic principle sounded familiar: solar radiation knocks electrons out of their specific atomic orbit, and electric fields between semiconductors help collect them in such a way that they become available as an electric current. The basic mechanism occurs at the nanometer scale; radiation is directly transformed into electricity, with heat produced only as a by-product of solar irradiation. There are no moving parts, and the atomic nucleus is not involved.

The initial efficiency of this conversion was 6 percent—about four times the efficiency of photosynthesis in plants when converting solar radiation into chemical energy. But the problem was how to manufacture large silicon areas of constant quality at acceptable prices. These problems were only solved later, when economic and political interests spurred serious development efforts.

However, essential advantages were obvious from the beginning: There were no specific dangers attached, so no unusual protection was needed.

Extreme Scalability Upwards and Downwards

The active layer of photovoltaic cells is nanometer-thin and has continuously decreased. All the rest is electrical contacts and packaging. The efficiency of the device does not depend on the size of the surface but only on the intrinsic properties of the specific technology chosen.

The necessary amount of active material was very small from the beginning and has decreased since. No unusual minerals are needed for mainstream silicon technologies. Silicon is available in abundance. Only relatively small amounts of silver, copper, or aluminum for the contacts need to be appropriately sourced and are worth being recovered after use.1635 Different types of thin-film-based cells use even less material, which in some cases is toxic but can be recycled.1636 Depending on the kind of installation, larger amounts of conventional glass, metal, and concrete may be needed for encapsulating the active layers and deploying the solar panels. New technologies for integrating an active photovoltaic layer into structural materials of buildings (roofs, facades, windows) or vehicles eliminate most encapsulation and all racking or mounting materials.

Because of these basic physical properties of the power generation process, PV is extremely modular, i.e., deployable at very different scales. Basic units are solar cells with a power rating of 5-6 Watt each. Very large cell factories produce up to two billion PV cells per year.1637 Cells are then encapsulated in modules (medium size ca. 1.8 x 1.0 meter) delivering 400–700 Wp (Watt peak). Such modules are used for all kinds of installations, from balconies to large ground-mounted power plants, although specialized sizes are now appearing. Further cost reductions through sophisticated mass production seem to be limited, but installation and balance-of-system continue to show major improvements. High modularity not only allows for extremely large production series but also for very rapid standard permitting and deployment.

High Innovation Rate and Strong Cost Reductions to Continue

The prevailing opinion is that the substantial cost reductions achieved by photovoltaics are primarily attributable to the advanced mass production techniques employed in China.1638 Such calculations usually refer to the module costs (per Wp). More relevant, however, are the resulting electricity production costs. They encompass important surface-related installation expenses that decrease with lower surface requirements when efficiency rises. Fuller analysis shows that gains in efficiency caused by advances in nanoscience-based material sciences are the main driver of solar electricity cost reductions.1639 While production costs per module area may not further decrease substantially, efficiency is expected to continue to improve considerably.

Learning curves illustrate how costs decrease with accumulated experience and over time. On a logarithmic scale, these curves can be approximated with straight lines. Figure 56 shows the learning curves in global weighted-average total cost of installed capacity for four technologies over the period 2010 to 2023. It illustrates that the costs of photovoltaics have come down much more rapidly than those of wind and CSP (a thermal technology). Advancements in nanoscale material science, on which PV technology essentially relies, have significantly contributed to this advantage and will likely continue to do so.

1. Learning Curves for Solar and Wind Technologies

Source: IRENA, 20241640

While the efficiency of silicon-based solar cells has increased beyond the expectations of only some years ago, new perovskite materials are approaching industrial maturity. These sophisticated materials are inexpensive and can be integrated into established production processes.1641 Changing the composition of solar cells can help tune the absorption spectrum. This would allow for the next big leap in efficiency: affordable multi-junction solar cells, where several layers are put one on top of the other to absorb different parts of the solar spectrum.1642 In the laboratory, tandem cells that combine silicon and perovskites have already achieved efficiencies of 34.9 percent.1643

Another development that may lead to substantially reduced solar electricity costs is the integration of an active PV layer into building and vehicle structures that are necessary anyway. Ongoing massive cost reductions of standard modules have hindered earlier attempts in this direction. But now, with accelerating PV deployment, new demand, new efficient thin-film materials, new encapsulation materials, and new methods for integrating inconspicuous rectangular cells, new approaches are emerging.1644

The Challenge: PV Transforms the Energy System, Brings a New Paradigm

Photovoltaic energy has the potential to provide electricity at a much lower cost than the predominant fossil fuel-based and nuclear power production methods currently employed on a global scale. However, it cannot simply substitute for conventional power plants. Within the present system, it faces a significant challenge: its ability to generate electricity depends upon the presence of sunlight or bright skies.

To overcome this limitation, solar power must be complemented by other energy sources, storage, and greater flexibility in the entire system, including consumption. Fortunately, new technologies appropriate to this purpose are evolving rapidly. The following pages will delineate the salient characteristics of these technologies. Together, they form an emerging renewable energy system, which will be further detailed in the following section (“The role of power generation technologies in energy systems”). This system operates differently from the old one and is driven by the increasing affordability of solar power, gradually leading to a new energy system paradigm.

New Batteries and Power Electronics

Neither new batteries nor power electronics are power generation technologies; rather, they are system technologies for the storage, transformation, and control of electricity. Together with photovoltaics and other distributed power generation technologies, such as wind power, they are about to deeply transform the whole energy system.

Both new batteries and power electronics, just as photovoltaics, have their origin in new material sciences enabled by quantum mechanics. As PV, they rely on processes in the electron shell of atoms, leaving the nucleus untouched. A bit later than microelectronics and photovoltaics, when advanced computing allowed for solving their intricate problems, they also experienced and still experience steep performance growth and cost decline.

Power Electronics—An Underestimated Key Component of the Energy Transition

Wind turbines, electric cars, photovoltaic devices, and efficient electric motors would not be possible or would have a hard time competing against traditional systems without modern, semiconductor-based power electronics. After the invention of the transistor, microelectronics specialized in treating digital information with ever smaller devices and minimal currents. Based on the same principles, power electronics were developed to control ever larger electric power flows. Using new materials and circuits able to deal with very large currents, voltages, and frequencies, power electronics learned to modify nearly every characteristic of power flows while steadily reducing power losses.

A key concept underlying power electronics is chopping up currents with semiconductor switches at very high frequencies and recomposing the parts to form currents with other characteristics. From tiny power supplies for everyday devices to house-high converters for high-voltage direct current power lines, from electric motors with variable speed to power control units in electric cars, power electronics has become widespread in all areas where electricity is needed.1645 Thanks to an improving understanding of the processes involved at the nanoscale and the development of new materials, the power density of power electronics devices has increased by a factor of 1,000 over the past 25 years.1646 And progress is not coming to a halt: presently, the stepwise substitution of predominant silicon with wide-bandgap semiconductors SiC (silicon carbide) and GaN (gallium nitride) is leading to considerable improvements.

Because of their common origin, digital data processing and communication technology can directly talk to power electronics without human intervention or old-fashioned electromechanical devices. Combined with innovative electronic sensors, this allows for far-reaching automation and robotics of all kinds. With these technologies, the traditionally “dumb” electrical grid, constructed with mechanical switches and heavy transformers for the top-down distribution of centrally produced power, is gradually being converted into a digitally controlled network that allows for highly efficient power flows in all directions, flexibly adapting to the demands and offers of millions of power producers and consumers.

New Batteries—A Game Changer for Flexibility

The most recent steep performance boost and cost decline among all major new energy technologies has been witnessed by rechargeable batteries, which are now starting to challenge conventional wisdom concerning the flexibility of power supply. Electrochemistry, allowing for the macroscopic transformation of chemical energy into electric energy and vice versa, has been developed since the early nineteenth century. But it was only with a deep quantum-mechanical understanding of the processes involved that a more powerful rechargeable battery based on lithium ions could be developed from the 1980s onwards.

The challenge consists of achieving high energy densities and rapid energy flows in multifaceted interactions between ions, electrons, complex hosting structures, and interfaces. The sheer multitude of potential combinations of chemical reactions, complex materials, and spatial configurations is such that important research efforts can lead to considerable progress in the field. The systematic use of extensive test series, large databases, and advanced quantum mechanical modeling has resulted in a significant acceleration of performance improvements. Known chemistries have been extensively refined, and new, promising combinations have been identified. Significant investments have been primarily driven by the aspiration to secure a substantial market share in the global battery market for electric vehicles. Nonetheless, batteries for stationary storage are also gaining importance. CATL, the global battery market leader, has over 12,000 staff members involved in R&D activities and holds 13,000 issued or pending patents worldwide.1647

Lithium-ion battery cell and battery pack prices have declined drastically over the past decade. Figure 57 represents the evolution of the volume-weighted average lithium-ion battery pack and cell price split over the years 2013 to 2024. The increase in 2022 was primarily due to a lithium shortage, which subsequently eased with the development of new lithium mines. The particularly strong price decline of 20 percent in 2024 was not expected and leads to an accelerated reevaluation of the role of batteries in the electricity system.

1. Lithium Battery Price Development, 2013–2024

Source: BloombergNEF, 20241648

Note BloombergNEF: Historical prices have been updated to reflect real 2024 dollars.

Today, global battery markets are dominated by various lithium-based chemistries, which are still being improved. In addition, a wide variety of new approaches, involving other elements or sophisticated nano-structured electrodes, promise a series of advantages:

• much higher power densities enabling farther reach of electric vehicles and even powering electric airplanes;
• cheaper, abundant basic materials, and no need for problematic, rare, or toxic resources;1649
• longer lifetime and lower costs;
• improved safety characteristics, new form factors, easier management, and cooling.

Given the numerous options, diverse approaches, and significant investments, further rapid progress in performance and reduction in cost is preordained. The potential for cost reductions through sophisticated mass production has also not yet been exhausted. This will have important consequences on all levels of the electricity system, from generation to consumption. Flexibility in time and space is increasing. As the section on system aspects demonstrates, significant price thresholds for widespread use in both public and private systems have been surpassed in 2024.

Game-changing Upgrades for Older Energy Technologies

Less spectacular progress has been made in other renewable energy sources and energy technologies that complement the disruptive dynamics of photovoltaics, new batteries, and power electronics. However, these technologies are essential components of the technology set that drives the emergence of a new energy system paradigm. As illustrated in the next section, together they can provide the flexibility needed to compensate for the least-cost but variable output of photovoltaics.

Wind Power’s Breakthrough with Power Electronics

Wind energy is essentially based on mechanical technologies. Increasing aerodynamic knowledge in the nineteenth century led to a shift from slow windmills to high-speed wind turbines, later to optimized blade forms, taking advantage of aerodynamic competence developed in aviation. From the 1950s onwards, a three-blade construction with horizontal axis emerged as the preferred solution. This design revealed two key parameters for further efficiency gains: increasing the hub height to reach higher wind speeds and expanding the rotor diameter to improve the relationship between the catchment area and blade length. As always in mechanical power plants, size matters. New materials, such as composite blade materials, and innovative tower construction technologies have enabled the development of present turbine sizes, step by step. The largest ones are not tolerated onshore, so they are being built offshore instead. This requires additional large amounts of steel and concrete for deep foundations or floating constructions. Moreover, construction and maintenance are much more complicated. Therefore, despite turbine size advantages and stronger and more frequent winds on the open sea, the costs per kilowatt-hour generated offshore are higher than onshore.

However, despite all mechanical improvements, a key problem remained: wind blows with variable speed, whereas traditional electricity generators needed to turn at a constant speed matching the frequency of the grid. Even the most elaborate gearboxes could not fully utilize the power of varying winds. Therefore, perhaps the most decisive enhancement in the efficiency and routine functionality of wind power plants, enabling the breakthrough of modern wind power, materialized on the electrical side. Since the late 1990s, advancements in power electronics have enabled the complete decoupling of generator speed from the frequency of the power delivered to the grid. Furthermore, digital controls facilitate the continuous optimization of the blade pitch, thereby ensuring maximum rotor efficiency in variable wind conditions.

Considering this background, wind energy’s cost decline has been understandably slower than in photovoltaics but still impressive. Moreover, further cost reductions seem difficult without a radically different mechanical design, which is not in sight. However, the cost per kilowatt-hour is not the only criterion for success. As shown in the next section, wind energy has important systemic advantages: the wind tends to blow stronger when the sun’s rays are weaker.

Meanwhile, new ideas are emerging in wave energy, a neighboring field which also aims to transform macroscopic natural mechanical energy into electric power. Here too, modern power electronics are enabling solutions that were previously unimaginable. Some nascent, still not industrialized technologies are claiming astonishing advantages compared to wind power.1650

The Electric Motor Becomes Flexible

The emerging energy system, increasingly based on solar and wind power, not only relies on new technologies for generating and storing electric power, but also on the massive electrification of energy use. Coupling transport and heat supply with the electricity system can lead to a strong increase in electricity demand and open up large new opportunities for energy efficiency gains and flexibility in the use of electricity. Heat accounts for almost half of global final energy consumption.1651 Transport consumes about one third.1652 New and improved technologies for converting electricity into heat and mechanical energy must therefore be considered an integral part of the set of technologies enabling renewable energy supply.

Traditional grid-connected electric motors operate at a constant speed since their functioning is linked to alternating current with grid frequency. Only with advanced power electronics has it become possible to regulate the speed and force of alternating-current electric motors by varying the frequency of the power supply. In many applications, using so-called Variable Frequency Drives (VFD, also called Adjustable Speed Drives or ASD) not only results in large energy savings, but also allows for much smoother operation and more flexibility in electricity consumption.1653 This is especially true for the half of electricity that drives pumps and fans, because many applications often require less flow than the maximum possible, and their energy use varies as the cube of flow, so (for example) halving the flow can save nearly seven-eighths of the energy.

The savings potential is impressive. In 2021, electric motors represented 46 percent of the E.U.’s electricity consumption. Power savings potentials for single units range 10–75 percent. However, as a VFD does not make sense for every application, the overall potential is estimated at 12 percent, of which only a quarter has already been implemented.1654 Secondary savings due to a more continuous operation, such as in Heating, Ventilation, and Air Conditioning (HVAC) systems, are not yet included in these figures. A fuller analysis, counting 35 improvements rather than just a few, found potential savings upwards of 50 percent.1655

Electric Heat Pumps Revolutionize Heat Supply Across All Sectors

Heat pumps are not new. Electromechanical chillers have been built since the mid-nineteenth century. The first ones for heating purposes appeared in Europe in the 1930s.1656 But as long as electricity was mainly generated with fossil fuels, they did not bring a real advantage. Today, heat pumps are becoming the key technology for electrifying heat use. They transport heat energy from one temperature level to another. The additional energy input necessary for this transfer depends on the temperature difference to be overcome and is much less than the heat energy transferred. Present technology heat pumps for smaller buildings in central Europe, sourcing ambient heat from the outside air, deliver on average three to four times more heat energy than the electric energy input they consume. For industrial purposes, large heat pumps today can deliver up to 300°C.1657 As for all plants using mechanical and thermal energy, the efficiency of compression heat pumps increases with size.

The recent boom in heat pump deployment in homes and industry, coupled with the continuous growth of the air conditioning and chilling markets, has led to a surge in research and development. Despite decades of experience with refrigeration, there is still much room for improvement. Efforts have mainly focused on substituting refrigerants that are harmful to the climate if released into the atmosphere and on reaching higher temperatures for industrial uses. Only recently have major heat pump manufacturers introduced VFDs for more efficient and flexible operation of home heating systems (“inverter heat pumps”).

However, more radical, partly nanoscience-based innovations are emerging. They avoid problematic refrigerants, achieve much higher efficiencies, produce heat of up to 400°C, or show a high dynamic flexibility in input and output temperature levels. One approach uses sound waves for high-frequency gas compression.1658 Another uses high-speed centrifuges in a condensation-less process.1659 Products using the latter approach are already available, and the market introduction of thermoacoustic heat pumps is imminent. The most fascinating approach, however, will still require some years of R&D to reach market maturity: electrocaloric solid-state heat pumps, in which a substance’s heat content varies with the applied electric voltage and heat pipes enable rapid one-way heat removal.1660 Even before the completion of a hypothetical nuclear power plant decided upon today, it can be expected that heat pumps will have 70 percent better efficiency, need much less space, be very flexible concerning temperatures, and need no refrigerants.1661

The availability of flexible heat pumps of various sizes for a wide range of temperatures enables the efficient exploitation of previously inaccessible heat sources. Waste heat from industrial plants, wastewater conduits in urban areas, rivers, and geothermal sources of all kinds can now be harnessed at varying temperatures at all levels. Traditional district heating systems needed to operate at temperatures above the final use temperature, resulting in significant heat losses despite extensive insulation. Today, heat pumps on the user side allow for fifth generation district heating and cooling, which operates at much lower temperatures. This enables district heating to avoid heat losses, deliver heat at varying temperatures, and even provide cold in the summer.1662 However, highly efficient buildings require so little heating and cooling that district heating systems may not compete with small, distributed, highly efficient heat pumps or even passive designs that do not need any active heating/cooling source at all.

Heat Storage May Solve the Winter Problem

The existing heat capacity of industrial plants, heating systems, and buildings already provides a welcome buffer and flexibility when connected to the electricity system. Every hot water tank is a heat storage device. In France, for example, even household-level tanks are mostly connected to the grid and thus adjustable. Various conventional storage systems have been developed for storage times of several days at different temperature levels. Furthermore, as the section devoted to systemic aspects demonstrates, heat storage over extended periods becomes crucial for compensating the fluctuations in solar power generation—at least for the inefficient, current majority of the building stock. Even before the widespread availability of heat pumps, so-called water Pit Thermal Energy Storage systems (PTES), i.e., very large, insulated water tanks, could deliver competitive heating energy during the winter when charged with heat from thermal solar collectors, generating most heat during the summer.1663

Eventually, these systems need to be much larger (for a better surface/volume relation) than the small systems needed for a single building’s heat requirements. Other technologies, not based on sensible heat (raising the temperature of a substance), can reach much higher densities and smaller heat losses, and thereby are interesting for installation in single buildings. As will become evident below, they may play a central role in balancing seasons in a fully renewable energy system. Phase-change storage systems operate at the melting point of substances (such as paraffin, ice, or salts) and store the melting heat required for the phase change from solid to liquid without changing temperature. Sorption heat storage (thermochemical storage) uses reversible processes where, in discharging, one substance (e.g., water vapor) brought in contact with another substance (e.g., porous ceramics, caustic soda) establishes weak chemical bonds while releasing heat, and in charging, heat input results in the separation of the two substances (by evaporation). Non-sensible heat storage usually comes at a higher cost if only considering the storage device itself—including the system costs of district heating grids required for large sensible heat storage, can change the picture. New developments in sorption technologies may offer disruptive opportunities for competitive seasonal heat storage at the building level within a few years.1664

Membranes for Industrial Substance Separation

There are also lesser-known recent technologies that strongly impact the energy system. In industry, flexibility and efficiency can be greatly enhanced by using innovative membrane technologies for separation processes, which today are accomplished with heat-based drying and distillation. According to estimates for the U.S., such technologies could save a quarter of industrial energy consumption. The special membranes required for this are not yet available for all processes. However, with currently known technologies, energy consumption in the chemical and food industry can be reduced by more than half.1665 In essence, macro-scale heating and cooling processes, as well as very large distillation columns that necessitate continuous operation in highly stable conditions, are being replaced with nano-scale physico-chemical selection processes in the membrane, driven by mechanical pressure. This substitution, therefore, not only results in higher energy efficiency, but also in considerably enhanced flexibility. Applied pressure can be varied in a certain range, pressure buffers with compressed air are cheap storage devices, and the units can be much smaller and shut off selectively. As industrial separation processes make up for around one-fifth of the world’s final energy consumption, membrane technologies can save up to 18 percent of global energy consumption and deeply flexibilize another two percent.1666

Hydrogen Technologies Suffer from Intrinsic Inefficiencies

Roots of the two key hydrogen technologies go back to the nineteenth century: electrolysis for hydrogen production with electricity and fuel cells for converting hydrogen back into electricity. These electrochemical technologies rely on the same family of principles as battery technologies. Therefore, the efficiencies of the core mechanisms do not depend on the scale. However, as material flows are involved, this is not true for an entire plant. In particular, electrolysis is more economical in larger units because additional parts like compressors, pumps, pipes, water treatment equipment, and sophisticated controls make up a significant portion of the overall costs.1667

The concept of employing electrically generated hydrogen as a substitute for fossil fuels was first proposed in the early 1970s. Important development efforts in hydrogen began in Japan. There were several waves of interest in hydrogen as an energy carrier, as it could be more easily stored than electricity.1668

While today’s lithium-ion batteries offer round-trip efficiencies (electricity to electricity) of more than 90 percent, electricity storage with hydrogen, even under optimal conditions, does not reach 35 percent, as losses in the individual steps accumulate: Electrolysis efficiency remains below 70 percent, and fuel cells or gas turbines can only achieve less than 60 percent efficiency.1669 Transport and storage losses, which strongly depend on distance and the overall system configuration, further reduce overall efficiency.1670 However, if heat use is also considered, in a combined heat and power setting, or simply by burning hydrogen, overall energy efficiency may reach 60 percent.

A particularly controversial topic is hydrogen transport costs. While the energy content of a kilogram of hydrogen is relatively high, the volumetric energy density (the energy contained in a certain volume at a certain pressure) is 3.3 times lower than that of natural gas.1671 Transport in pipelines, therefore, requires larger pipe diameters and much stronger pumps. Moreover, many materials used in traditional gas pipeline technology become brittle when exposed to hydrogen. The extent to which existing gas pipelines can be repurposed for hydrogen transport and the associated costs remain controversial. In comparison, transporting hydrogen on ships for long distances, necessitates liquefying it by cooling it down to -253°C (20 Kelvin). At these temperatures, evaporation (boiling off) of a certain share of the cargo is inevitable. Liquefaction and regasification of hydrogen are energy- and capital-intensive processes.1672 Mostly underestimated are the negative climate effects of hydrogen losses.1673 Many ambitious hydrogen strategies, such as those promoted by Germany, the E.U., or Japan, have included plans for extensive international trade with green hydrogen, taking advantage of low electricity generation costs in countries with abundant sunshine. Because of the losses and costs involved, the idea of converting green hydrogen into ammonia or more sophisticated e-fuels before transporting them over long distances is gaining traction as a more realistic solution.

The attraction of hydrogen for politics and incumbent industries lies in the hope of having found a clean drop-in substitute for fossil fuels requiring little technical modifications. However, expectations have dwindled over the past two years. Cost projections have proven to be too optimistic, demand for green or pink hydrogen (electrolysis with renewable or nuclear energy) has not developed as forecasted, and huge investments into public infrastructure, including import facilities and distribution networks, are increasingly being questioned. Nevertheless, using clean hydrogen as an energy carrier in the overall energy system is not its only possible application. With the decarbonization of the chemical industry, demand for hydrogen as a material non-energy feedstock will increase.1674

Conclusions

Table 17 gives an overview of the main characteristics of relevant energy technologies discussed in this section.

1. Key Characteristics of Main Energy Technologies Promoted for Decarbonization

Key conclusions from this comparative analysis can be summarized as follows:

• Atomic energy is the only technology exploiting nuclear forces. It therefore has safety and security problems, as well as long-term liabilities, that competing technologies do not have. The consequences are rising costs and wastes to be safely stored for millennia.
• Technologies that exploit nano-level mechanisms in the electron shell of atoms, aided by quantum-mechanical insights, exhibit the most rapid innovation and cost decline.
• For generating electricity, photovoltaics has the most attractive properties in nearly all categories and the lowest electricity generation costs. However, it is not a drop-in solution for substituting conventional power plants; it needs a set of complementary technologies providing flexibility.
• Such complementary technologies have developed over the past decades with increasing speed. The most important ones are power electronics, batteries, and wind energy. Not only do they complement solar power in the electricity sector—they have started to deeply transform and decarbonize the whole energy system.
• Material sciences, powered by a combination of quantum mechanics and increasing computer performance, including AI, enable dramatic improvements in the performance of new and more conventional technologies. Highly efficient conversion, storage, and control technologies enable energy flows that previously have not been economical. This allows for more complexity and flexibility in the system.
• Solar power, therefore, or in a larger sense, renewable energy, is not about simply substituting one energy source for another. It fundamentally changes the way of conceiving energy systems. A new paradigm is already in the course of implementation.

New Roles for Energy Technologies in a Fundamentally Changing Energy System

Powerful Drivers of Change

The previous section on Radical Changes in the Context of the Commercial Use of Nuclear Power has provided an overview of the changing landscape conditioning the development of nuclear power. This section will analyze in more detail how the technology developments outlined in the last section are changing the electricity system, in particular in the following areas:

• generating electricity from renewable sources
• converting between different forms of energy
• controlling energy flows
• storing electricity and heat

The two main drivers of this change are the need for decarbonization and technological progress, which are intrinsically intertwined. The perceived necessity to stop human-made climate change by ending the use of fossil fuels has triggered a wide range of political and industrial initiatives. These initiatives have started or accelerated technological developments and changed markets and cost structures. The result is a set of powerful, albeit complex, economic dynamics.

New technologies are profoundly changing not only electricity generation and electricity distribution, but also the role of electricity in the overall energy system. Slightly reframed, the main drivers of this role change are important and rapid improvements in performance and cost in the specific functions of digital flows control and complex interactions.

The enhanced opportunities for converting between different forms of energy, together with the need for decarbonization, are strongly encouraging the coupling of various sectors of the energy system that previously operated rather independently. In the emerging, more complex system, electricity is acquiring a central role. It is the most universally convertible form of energy.1675

Power Generation: Renewable Electricity Unbeatably Cheap

Basic physico-technical considerations and the development of deployment statistics both show very clearly that, in spite of headwinds in some countries, photovoltaics and wind will almost certainly dominate power generation in the next decades.

As the analysis in the last section has shown, photovoltaics stand out as the most cost-effective, least harmful, and least disturbing new clean energy source with the highest innovation and deployment potential. Roof surfaces alone would be sufficient for covering 1.8 times the present electricity consumption.1676 Moreover, it presents an attractive investment opportunity at all scales that can be deployed easily and rapidly, including on small buildings behind the meter. This latter characteristic significantly expands the pool of potential investors and operators, and, as will become more evident below, changes the role of the public grid. However, photovoltaics does have an important weakness: its fluctuating nature necessitates flexible complementary technologies.

At the level of power generation, wind is the most important complement. The global potential of wind energy is theoretically more than sufficient for covering all energy needs. However, near settlements and in beautiful landscapes, wind energy is a disturbance and often meets local opposition. Offshore wind energy is more expensive and needs longer connections.

The potential of other renewable power generation sources is limited. Large hydro energy (dams and rivers), which has its own environmental and sustainability issues but should nevertheless be mentioned as an alternative, requires appropriate geographical structures. In most OECD countries, the potential has largely been exhausted. Moreover, hydrological regimes are becoming unreliable due to climate change.1677 Bioenergy use has been strongly extended in the past decades. However, its overall efficiency is low,1678 and bioenergy can pose additional problems like displacing natural habitats, competing with food production for surfaces and irrigation, being exposed to climate change affected agricultural yields, and, last but not least, it will be needed for the substitution of fossil resources in the chemical industry.1679 Both hydro energy and biomass, as well as waste, have the advantage of being, to a large extent, dispatchable electricity sources, i.e., their output can be actively adjusted according to needs. Geothermal power production requires high temperatures and usually deep drilling. Areas where these can be accessed with reasonable effort are restricted. Capital costs are high and require rather continuous operation. Shallow geothermal energy, on the contrary, as a lower temperature heat source or as a storage medium, is accessible nearly everywhere and can become important for heat supply, especially in conjunction with heat pumps.1680

The most recent estimates of the Levelized Costs of Energy (LCOE), i.e., the full costs per kilowatt-hour over the lifetime of a facility, not including distribution costs and not considering the temporal availability, show a clear advantage for PV (see Figure 58).

In a detailed study for Germany, Fraunhofer ISE concluded that in 2024, utility-scale PV plants in southern Germany could produce at €c4.1/kWh (US$c4.5/kWh), whereas small rooftop installations, delivering at the retail level, at €c6.3/kWh (US$c6.8/kWh). Compared at the retail level, PV rooftop electricity costs were lower than the costs of PV utility scale plus distribution. Onshore wind power came at €c4.3/kWh (US$c4.7/kWh) for good locations. Using moderate assumptions for learning rates (PV 15 percent, wind 5 percent), Fraunhofer estimates that PV in 2045 will decrease to €c1.9/kWh (US$c2/kWh) (ground-mounted) and €c4.9/kWh (US$c5.3/kWh) (rooftop), while good wind locations would reach €c3.7/kWh (US$c4/kWh).1681 For the MENA region, ISE estimated best values of €c5.3/kWh (US$c4.7/kWh) for small rooftop and €c3.5€/kWh (US$c3.8/kWh) for ground-mounted systems in 2024. All amounts expressed in US$2024.

1. Levelized Cost of Energy for Various Energy Technologies at Different Locations in Germany

Source: Frauhnofer ISE, 20241682

Notes: CCGT-CH4: methane combined-cycle gas turbine; GT-CH4: methane gas turbine.
The Figure represents LCOEs of renewable energy technologies and conventional power plants at various locations in Germany in 2024. Obviously, nuclear power plants are not in operation in Germany anymore. Specific system costs are considered with a minimum and a maximum value per technology.

International studies need to estimate multi-country averages. The samples can be constructed in different ways, resulting in slightly differing averages as well. Moreover, LCOE formulas used may not be exactly the same. According to the 2025 edition of Lazard’s widely referenced report, the estimated lower values for wind onshore are US$c3.7/kWh, for PV utility US$c3.8/kWh, for wind offshore US$c7/kWh, and for PV community & CI US$c8.1/kWh. For U.S. nuclear, the lower LCOE is US$c14.1/kWh (3.7 times as much as utility solar), and for gas peakers US$c14.9/kWh. With US$c3.4/kWh, purely operational costs for nuclear power plants are in the range of wind onshore and PV utility.1683 Saudi Arabia, for comparison, reported an LCOE of US$c1/kWh for utility solar in June 2024.1684

In addition to LCOE calculations based on the technologies themselves, there are different approaches to quantify the integration costs into a specific (present or future) energy system, such as the Value Adjusted LCOE (VALCOE) of the IEA and the firmed renewable energy LCOEs of Lazard. As they depend on the specifics of the energy system considered, they may strongly differ between contexts.1685 Moreover, the values depend on the method chosen. The extra cost of firming can range from high for 100 percent battery backup to less than zero for battery-free demand-side grid-stabilizing resources. Not to forget, thermal power plants with their large, sharp, and often unpredictable outages also have firming costs.1686

The fluctuations of solar radiation due to astronomy (daily and seasonal variations) are regular and fully predictable. As demonstrated below, the 24-hour day/night cycle can be fully compensated by today’s batteries at a reasonable cost. For the 12-month seasonal cycle, which is more pronounced in latitudes farther from the equator, this is not a viable option. These regular cycles are superimposed by weather effects that can last for several days and are governed by stochastic processes. Wind has no daily or seasonal astronomic cycles; moreover, for periods between one and 24 hours, forecasting errors for wind are larger than for PV.1687 It tends to blow stronger when solar radiance is less intense.

However, depending on geography, wind power may even fully compensate for the seasonal variations of solar power. Elia Group, the transmission grid operator in Belgium and eastern Germany, writes in its 2021 vision for a climate-neutral European energy system: “Long-term fluctuations can be minimised with a good balance between wind (generating more energy in winter) and solar energy (generating more energy in summer), avoiding a seasonal mismatch between power supply and demand.” 1688

The distinction between utility-scale and rooftop photovoltaics is essential. Rooftop installations are usually owned by the building owner or a community of building users. They are connected behind the meter, they are not part of the public electricity system as long as the electricity is consumed in the building itself, and grid fees and (in most countries) taxes do not apply to them. This fundamentally changes the economics. Worldwide, in 2023, according to the IEA, 44 percent of the cumulative installed PV capacity was of this type; in Germany, it reached 78 percent.1689 In most countries, privately installed PV panels behind the meter and/or electricity produced for self-consumption are not properly accounted for in public commercial energy statistics.

As behind-the-meter installations continue to scale, their capacity is becoming increasingly important. This was recently illustrated in Pakistan: The 10-percent decline in power demand in the public grid between 2022 and 2023, while the economy grew, can be partly explained by examining the import statistics for PV panels. In 2024 alone, the imported solar generation capacity amounted to 22 GW (compared to 46 GW of mostly conventional capacity operating in the public grid in 2023), primarily adding to private installations behind the meter.1690 The chaotic development, intensified by sharply rising public electricity prices and unsatisfactory reliability of the grid, is creating new opportunities but also causing social problems. It highlights new challenges in managing the transition in emerging economies with intense sunshine.1691 In the case of Pakistan, it appears that a combination of key cost trends has reached a systemic tipping point.

Electricity Use: Electrification Brings Efficiency and Flexibility

Sector Coupling Expands and Stabilizes the Electric System

Electrification of the heat and transport sectors is a key strategy for decarbonization and an official policy goal in most countries. While this means more electricity will be needed, it also opens up opportunities for greater temporal flexibility in power demand due to the inclusion of more use cases.

Often, there is confusion between the installed generation capacity (measured in megawatts or gigawatts) needed to cover consumption peaks on one hand, and the overall electricity demand over a day or a year (measured in megawatt-hours, i.e., a megawatt power for the duration of an hour) on the other hand. As flexibility can reduce peak demand, increased flexibility can reduce the need for peak generation capacity. However, because solar and wind power only produce when the sun shines or the wind blows, they usually need more generation capacity than nuclear power to produce the same amount of electricity.

Considering the energy (MWh) needed, it is important to understand that the electrified overall energy system requires more electricity, but overall much less energy than the old system. As we have seen in the previous section, electrification offers many opportunities for higher efficiency. Energy units presently consumed in the form of fossil fuels for heat and transport purposes will therefore translate into far fewer energy units in the form of electricity. Two main factors contribute to this reduction: the electrification of propulsion (substitution of combustion engines with electric motors) and the electrification of heat production (substitution of combustion with heat pumps). While combustion engines lose about two-thirds of the energy input as heat, electric motors have efficiencies of over 90 percent. Heat pumps, on the other hand, can harness about two-thirds of their heat output from ambient heat. As a result, the system reduces its energy losses in propulsion and gains additional energy in heating.

Flexibility: Untapped Potential on Multiple Levels

While estimating overall efficiency improvements can be challenging, the concept of efficiency is relatively straightforward. Efficiency is defined as the ratio of output to input energy. For a specific service or product, historical data allows for comparisons.

The concept of flexibility, on the other hand, refers to the potential for changes in operational modes. It must at least include speed, frequency, and duration of possible changes compared to a standard. This multitude of dimensions makes it difficult to compare costs across technologies. Historical data are usually not available because the technical limitations, consequences, or costs of such changes have mostly never been tested, as nobody has requested it. In most domains, market designs did not incentivize flexibility, and control options were missing.1692

There is a wide array of opportunities for flexibility in the electricity system. It ranges from reserve generation capacity provided by gas peaker plants to excess solar generation capacities curtailed during maximum sunshine, from batteries in homes to intermediate heat storage in industrial plants, and from changes in consumer habits or cost-neutral power use patterns to changes in industrial production rhythm enabled by additional intermediate product storage.

For the short-term management of electricity grids, a series of flexibility “products” have been defined that already trade on specific markets. These products include balancing power with different reaction times as well as other ancillary services linked to the specificities of three-phase alternating current. However, access to these markets has historically been limited to large operators, and their spatial reach is a subject of debate. (See also Figure 68 in WNISR2024).

An important compensation mechanism for up to ten seconds is provided for free by the inertia of the large rotating masses of the turbines in large thermal or hydro power plants. Some are still arguing that this is a reason for maintaining such power plants. However, the technology of inverters that in solar or wind power plants convert direct current (DC) into alternating current (AC) with grid frequency has evolved considerably. In addition to “grid following” inverters that adapt to grid conditions, there are now “grid forming” inverters that can take over the role of rotating masses. In order to stabilize the grid in critical situations, however, there needs to be a sufficient share of gird forming inverters, and they require coordination. “Virtual inertia” can be faster, more precise, and more effective than the mechanical inertia of classical rotating machines. The much-discussed blackout on the Iberian Peninsula in April 2025, which immediately triggered calls for not giving up thermal power plants with rotating masses, was caused by an insufficient adaptation of the grid to renewable energy sources and a lack of balance in the deployment of these new technologies, as well as failure of thermal plants to deliver the voltage control they had agreed and were being paid to provide.1693 How to precisely define the requirements for inverters is an ongoing learning process.

In the medium- and long-term, however, there will be a much easier way to guarantee a very stable functioning of electricity networks with millions of active participants: the transition of ever larger parts of the grid from three-phase alternating current, with its complicated phase-shift parameters, to direct current, where voltage, and not frequency, is the main coordinating reference. With the development of power electronics, more and more AC-based traditional transformers and ancillary equipment can be replaced by more flexible semiconductor-based devices that can convert direct current across all voltage levels and switch it safely.1694 In a conservative and safety-oriented business, however, where important investments are based on well-established standards, such a transition takes time and meets resistance.1695

While the flexibility of electricity generation has always been a key issue, the flexibility of energy demand has gained much more interest only recently. Timid experiments with opening access to traditional flexibility markets, with additional flexibility products such as grid capacities, and with more differentiated spatial reach of these markets (zonal or nodal pricing) have shown that flexibility margins are larger than previously estimated and that they have grown over time as appropriate investments were made.1696 Only real-world experience can show the potential under specific conditions. Available literature, therefore, often refers to monetization options in specific market designs, which, however, will undoubtedly change, making general statements and comparisons difficult.

In the early 2000s, it was widely believed among experts and often repeated by utility representatives that the electricity system could not manage to have more than 10 to 20 percent of variable generation from wind and solar.1697 Meanwhile, confronted with the pressure to integrate fluctuating renewables, much more flexibility has been mobilized. Today, in many systems, variable renewable sources make up more than 50 percent of total generation for periods of at least several days. Several European countries with modest or no hydropower get more than half their annual electricity from renewables, while the GW-scale grid of South Australia, with zero “baseload” capacity, was 74 percent powered by variable renewables in 2024, with its grid operator forecasting 100 percent by 2027.1698 The wide and confusing variety of flexibility options raises the question of why they have not been discovered and utilized earlier.

Various flexibility mechanisms have long been employed, either as ingrained practices or when circumstances required them. In Italy, for instance, standard household grid connections have traditionally been capped at 3 kW. It was widely understood that using too many appliances simultaneously would trip the automatic circuit breaker. In the early 2000s, international observers were surprised by how quickly Italy’s national electricity provider deployed electronic meters in nearly every household—an undertaking that proceeded much more slowly in other countries. The utility’s motivation was clear: replacing conventional fuses with electronic switches significantly reduced the margin of error in power control. This, in turn, curbed peak electricity demand, eliminating the need to build several new power plants. The savings generated from avoiding those investments were used to fund the full-scale rollout of digital metering instead. Additionally, the new meters paved the way for introducing multiple time-of-use pricing structures.1699

In France, the need for demand-side flexibility emerged earlier than elsewhere because the share of rather inflexible nuclear power was higher than anywhere else.1700 Initially, compensating for the inflexibility of nuclear power was easy, since the building up of the electricity system in France was to a very high extent based on hydro power, which is highly flexible and easy to regulate. Since EDF owns both nuclear and hydro power plants, costs for this flexibility were not made explicit. Promoting the use of electricity for heating and hot water supply with special tariffs was then an easy way for early forms of demand-side flexibility. In 2020, 37 percent of homes had electric space heating, of which 5 percent used heat pumps. As a result, the difference between day and night consumption in France is lower than in other countries. However, increasing differences between winter and summer consumption limited the use of this approach.1701

In many countries, however, utilities were reluctant to pave the way for solar power, and changing traditional habits and expectations of households or industrial users is a long-term task. An important reason for using only a small selection of the above-mentioned flexibility options is the difficulty of coordinating them. Yet this becomes possible with modern power electronics and appropriate market designs. A more detailed discussion of this aspect follows below.

The expansion of the electricity system into the heat and transport sectors, which widens the range of usage patterns involved, considerably enhances the opportunities for flexibility. Moreover, the section on the physical principles has shown how new technologies have widened flexibility options in many regards. Additional storage is not always the most economical option. However, spectacular advances in electrical storage and the new role of heat in the overall system make it worthwhile to analyze their impact in more detail.

Electricity Storage: Battery Costs Reach a Tipping Point for the System

In 2024, the price of lithium-ion battery packs declined by 20 percent, reaching US$115/kWh towards the end of the year. The corresponding cell price was US$78/kWh.1702 This was not simply due to dumping prices: according to the Volta Foundation, the leading international battery expert organization, the average LFP battery cell production cost in China by Tier 1 producers in 2024 was US$60/kWh. BYD, the second largest producer, achieved a cost of US$43/kWh. Volta estimates that an average of only US$25/kWh is on the horizon.1703

This has worldwide consequences. Retail prices for ready-to-use home storage kits range from US$200 to over US$1,000/kWh. A simple rule-of-thumb calculation based on actual offers for two popular brands, BYD and GLCE, yields levelized costs of storage of US$c8.9/kWh and US$c3.4/kWh, respectively.1704 Assuming that 100 percent of the electricity generated by rooftop PV is being stored to be delivered at any time during the following 24 hours (just for setting an upper boundary), taking Fraunhofer’s lower values for rooftop PV in southern Germany quoted earlier and applying the two indicated levelized cost estimates leads to an overall cost for solar power delivered around the clock of US$c15.2/kWh (or even US$c9.7/kWh)—about half of the present utility power retail price in southern Germany.1705 In reality, for households and small businesses, the percentage of own electricity produced and consumed that needs storage will be much lower. This means that the incentive for installing storage systems that level out 24 hours is even higher.

The effect of this rapid development is also being felt on the other side of the scale. In China, the market price for large grid-connected storage systems (completion in 2025 or 2026) is now US$66 per kWh of capacity.1706 The price, which was set in an auction that ended in December 2024, included maintenance services for 20 years. This means the cost of storing energy at this scale in China is about US$c1/kWh.

The price decline has led to a surge in installations over the past three years. In Europe, the cumulative storage capacity has grown by an impressive 58 percent annually (see Figure 59). Two-thirds of these European installations are behind the meter, demonstrating the influence of private economic considerations.1707 Utility-scale BESS (Battery Energy Storage Systems) have been hindered by (technical or bureaucratic) connection problems to the grid—in many countries, waiting queues for connection are long.1708

At the global level, the average annual growth of cumulative capacity over the same period has even amounted to 80 percent. Countries leading this deployment were China and the United States. In 2024, Europe accounted for only 17 percent of global capacities (61 GWh of 352 GWh).1709 Volta Foundation’s estimation of a mere 20 percent of installations behind the meter globally appears to be an underestimation when considering the proportions in Europe and the poor accounting of behind-the-meter solar installations in many countries.

1. Grid-Size Battery Fleet in the EU27, U.K., and Switzerland, 2015–2024

Source: SolarPower Europe, 20251710

Note: *EU27, U.K., Switzerland.

While the share of overall battery production being used for BESS has increased from 7 percent in 2020 to 15 percent in 2024, the vast majority of batteries are still being used in electric cars.1711 The success of policy-driven strategies for pushing e-mobility adoption has fallen short of expectations. Battery development, mainly driven by the Chinese government and Chinese companies, however, has pushed cost and performance to such a point that cost arguments in favor of electric cars are becoming increasingly compelling. Within ten years, electrification in the transport sector will therefore have a massive implications for the energy system.

Private cars are sitting idle on average for about 90–96 percent of the time. From an energy system perspective, their batteries are underutilized and could be an extremely valuable asset if connected to the grid when not in use for driving (bidirectional charging). In all scenarios, the charging of large numbers of electric cars must be distributed over the day by some coordinating mechanism in order to avoid a problematic and expensive surge in peak demand. If all present cars in Germany were electrified according to current standards, their combined capacity would equal two days of the country’s current electricity consumption.1712 A loss in service comfort and a shortening of battery lifetime through bidirectional charging is not to be feared: In Germany, a fully charged standard car battery lasts for eleven days of average driving, the possible number of charging cycles is far from being exhausted over the lifetime of an average car.1713

A fully charged standard car battery lasts for eleven days of average driving, the possible number of charging cycles is far from being exhausted over the lifetime of an average car.

The costs and performance of stationary battery storage systems have reached a tipping point, making their widespread use in the electricity system a compelling prospect. They have achieved a level of proficiency that enables them to balance the fluctuations in renewable power generation over a 24-hour period, both from a technical and economic standpoint. This has profound consequences for the economic viability of renewable power generation, especially solar. Moreover, equalizing electricity flows over 24 hours, significantly reduces the need for grid expansion, as discussed below.

While these calculations referred to 24-hour cycles, the economics of storage look totally different for long periods of low renewable power production, especially seasonal variations of sunshine. If used for only one or two cycles per year, the investment cost of lithium-ion batteries with a limited lifetime of 20, perhaps 30 years, would lead to very high costs per kWh actually stored1714—for strong seasonal cycles of fluctuating renewable power, other storage technologies are definitely welcome. Heat storage and hydrogen are two main options.

Heat Storage: The Ideal Buffer for a Large Part of Electricity Consumption

While the electrification of the transport sector causes increasing electricity demand, it can add flexibility to a significant extent, as shown above.

The electrification of heat production has more complex consequences. Implemented straightforwardly, it would exacerbate seasonal differences: In regions far from the equator, heat demand is higher in the winter, but solar power generation is stronger in the summer. In other regions, power demand peaks in the summer because of cooling needs and thus constitutes a better match for solar PV.

An interesting solution to cover the winter needs could be long-term heat storage. This would allow for the generation of heat with high heat-pump efficiencies and abundant solar electricity in the summer for use in the winter. This would increase efficiency by shifting considerable amounts of electricity demand from the winter to the summer and add considerable flexibility to electricity consumption since producing heat for storage would not be time-sensitive.

Existing competitive seasonal heat-storage technologies require large plants feeding local heating networks. Their centralized heat generation adds flexibility to the public electricity grid.

In June 2025, the largest sand battery in the world started operating. It can deliver 1 MWh of thermal power and store up to 100 MWh of excess energy from variable renewables. It can feed the local urban heating system or provide 400–500°C hot process heat for the industry.1715

Advanced sorption heat storage, on the other hand, is not yet competitive, but it will likely be on the market before new nuclear power plants can provide power. This technology could be installed in a single building and fed with heat produced by solar power generated behind the meter. It would considerably enhance the competitiveness of integrated energy generation and energy use by private prosumers (system participants that are both producers and consumers of energy and can balance production and consumption behind the meter for much of the time). The technology’s ease of installation, permitting, and financing suggests a potential for rapid growth, similar to the current trend in home electricity storage.1716

Hydrogen and E-Fuels: Costly Workarounds for the Transition

Many scenarios for the future energy system include the use of gas peaker plants fueled with green hydrogen. As hydrogen can be stored in caverns for many months, this seems to be an obvious solution for seasonal storage. Countries such as Japan and Germany, as well as the E.U., have developed ambitious plans for a hydrogen economy. However, as round-trip efficiencies (electricity–hydrogen–electricity) will not exceed 35 percent, this will remain a rather expensive solution requiring high investments in complex infrastructures.1717 As the expansion of renewable electricity production is a bottleneck in the overall system transformation, green hydrogen production may not only be a waste of capital, but also lead to a diversion of badly needed green electricity and thereby slow down the transformation (see previous WNISR editions). From a system efficiency perspective, a combination of distributed long-term heat storage and other flexibility options is a more elegant and cost-effective way of balancing medium- to long-term electricity shortages. However, they require more advanced market designs and bottom-up control concepts than those in place today. If gas peakers are considered to be inevitable in a transition phase, the use of green ammonia as a storage medium and fuel may be preferable to green hydrogen. While green ammonia’s production requires green hydrogen in the first place, and is therefore more expensive at the source, its handling, transport, and storage are much easier and necessitate much smaller infrastructure investments.1718 Ammonia is a traded commodity on the global market because it is an essential chemical industry feedstock, particularly for the production of fertilizers.

Ammonia is but one example of e-fuels, which are fuels produced with carbon-neutral electricity. They are being discussed not only for electricity generation at demand peaks, but also as fuels for other hard-to-abate purposes, mainly maritime transport and aviation. There is a wide variety of scenarios analyzing different e-fuels, different local production and import strategies, and different uses.1719 None of them, however, can deny the fact that at least 40 percent of the energy input in the form of electricity is lost in the process of producing the e-fuel, and—except for heat uses—another 45 to 65 percent is lost in the engine transforming fuel into motion or electricity. Transport and handling losses come on top. When direct use of electricity or more efficient storage alternatives are conceivable in the medium term, infrastructure costs and lock-in effects have to be considered. Even for aviation, one of the hardest-to-abate sectors, electrification is extremely desirable as battery-driven airplanes would need around one quarter of the electricity of those powered by e-fuels.1720

In any case, hydrogen and e-fuels will be crucial material feedstocks for decarbonizing the chemical industry.1721 Therefore, using surplus electricity, when not needed elsewhere, to produce hydrogen as a byproduct of the energy system makes perfect sense. However, using hydrogen and e-fuels as energy carriers will remain a costly exception. It may be inevitable in transitory phases, but it does not make sense in the long term.

System Integration: Enhanced Interlinkages Change the Control Paradigm

The increasingly central role of electricity in the energy system and the tighter coupling of different sectors call for more attention to the character of the growing number of interlinkages and their consequences.

Power Electronics (PE) enable bidirectional flows and nearly lossless conversion (AC/DC, frequency, voltage, phase shift etc.). Combined with Information and Communication Technologies (ICT), PE is increasingly allowing for full digital control of electricity grids. More generally, power electronics, ICT, and artificial intelligence together enable fully digitalized and automated control of generation, consumption, and flows from the top level of the system down to the household level. This opens a wide range of scenarios involving interactions between millions of subsystems and actors.1722

Traditional electricity systems were controlled by adapting the power generation of a limited number of large power plants to electricity demand. Load imbalances were indicated by deviations of the grid frequency. Generous safety margins for the tricky parameters of three-phase current enabled the safe operation of essentially “dumb” grids.

When combined with decentralized storage at various grid levels, optimization of power parameters at each grid node can significantly increase the capacity of existing grids, especially in the distribution network. Electricity flows can be distributed over time, and traditional security margins in electrical parameters can be reduced.

Furthermore, the surge in photovoltaic and storage installations behind the meter means that a growing share of overall electricity production is consumed onsite rather than flowing through the public grid. Sophisticated energy management systems in homes, buildings, and industries can optimize energy flows behind the meter according to a wide range of criteria and can control the exchange of energy with the grid.1723

But with the need for more flexibility to accommodate variable renewable power generation, the number of assets that generate, consume, or store electricity and need to be coordinated in some way is exploding. Millions of owners and decision-makers are involved, some delegating short-term decisions to automated management systems while occasionally changing preferences, such as the time they need a fully charged car. Meanwhile, incumbent utilities’ attempts to maintain central control by directly controlling major appliances behind the meter are failing due to complexity, social acceptance, or falling short of realizing even a small share of the potential. Increasingly, market mechanisms are being explored or used for coordinating individual decisions. In Germany, where wind and solar power accounted for 47 percent of electricity consumption in the public grid in 2024, discussions on market design have intensified. Detailed expert papers from a diverse range of specialized think tanks are contributing to this debate.1724 Sophisticated energy management systems can optimize energy exchange across the meter in reaction to price signals, weather forecasts, and behavior patterns.1725 Similarly, transactions between grid operators, power plant operators, storage operators, system integrators (“virtual power plants”), and prosumers are increasingly coordinated through appropriately constructed markets for power products, grid capacities, or specific ancillary services.1726

Special agencies in countries or trading zones are busy designing new markets according to perceived new necessities following changes in the technology mix. Often demands and interests of old and new participants in the energy system are diverging and clashing, but the overall tendency is clear: There is a consistent shift in the control philosophy of the system—from centralized control in a hierarchical system with tendentially monopolistic structures to bottom-up control in a complex network that needs carefully designed markets to set the right incentives. The intrinsically monopolistic infrastructure of the physical public grid must be kept open, functioning, and optimally utilized by a wide variety of private actors.

All these trends, including the transitions from

• fossil fuels to electricity,
• centralized to distributed power generation,
• dispatchable to independently fluctuating power sources,
• careless energy consumption to price-sensitive flexibility,
• public supply to autonomous prosumers,
• monopolies to complex networks of millions of actors,
• central planning to complex market interactions,
• top-down to bottom-up control,

constitute a profound paradigm shift that necessitates intensive learning and adaptation on all levels.

Consequences of Long-Term Trends:
Key Characteristics of the Coming Energy System

In many countries, the trends described here are not yet dominant. However, by 2035, when a hypothetical new nuclear power plant currently under consideration could potentially become operational, and halfway to the political goal in most countries to fully decarbonize by 2050, these trends will have profoundly transformed our energy systems.1727

The emerging energy system can be characterized as follows:

• PV and wind will dominate power production.
• Electricity will play a central role in the energy system as the main energy source (generated with sunlight and wind) and as a universally convertible energy carrier.
• Flexibility will be a highly appreciated characteristic of all assets in the energy system.
• Market mechanisms, mediated by digital controls and power electronics, will play a key role in coordinating millions of actors.
• A consistent part of the energy used will be generated behind the meter and not pass through the public grid.
• Residential and service buildings will deliver power to the grid in the summer and withdraw power from it in the winter, while being able to react to market prices in the short term.
• Consumption flexibility, battery storage, and heat storage may be complemented with combustion of green hydrogen. E-fuel use for aviation will be difficult to avoid in the short term.
• Carbon-neutral hydrogen production with excess power will be essential for non-energy uses in a decarbonized chemical industry.
• Low-income countries with high solar radiation and low seasonal variations will likely leapfrog the development of conventional “dumb” grids with central power stations.1728

Integrating Nuclear Power into the Energy System: Increasing Difficulties

In the above analysis of the energy system’s ongoing transformation, it has already become clear where and how nuclear power has a different logic and faces growing difficulties in integrating into the increasingly dominant system. Nevertheless, supporters claim that there are new opportunities for nuclear energy. The following pages will take a closer look at these arguments while summarizing the challenges. As a starting point for analyzing possible future roles of nuclear energy, it is useful to briefly look at the role for which it was originally conceived.

Nuclear Power in Conventional Energy Systems

The commercial or civil use of nuclear power was born as a more likable version of a radically new technology developed for military purposes.1729

In the 1960s and 1970s, the heyday of atomic energy, nuclear power plants were meant to simply complement or substitute large thermal power plants. In 1980, at the apex of reactor constructions, the global electricity generation mix was dominated by coal (38 percent), oil (20 percent), natural gas (12 percent), hydropower (21 percent), and nuclear energy (9 percent), with non-hydro renewables accounting for less than 1 percent.1730

In the year 2000, nearly two decades after the early wave of nuclear power plant construction, only the following countries generated more than 50 percent of their electricity with nuclear: France (76 percent), Lithuania (74 percent), Belgium (57 percent), and Slovakia (53 percent).1731 All of these were integrated into larger interconnected regional networks, which allowed them to comfortably run their nuclear plants at maximum capacity, as fossil fuel plants with lower fixed costs would buffer demand fluctuations. Fossil fuel plants would also compensate for planned and unplanned outages of nuclear reactors. Among the other countries on the list with a high reliance on nuclear energy, only South Korea (41 percent of electricity provided by 16 commercial reactors) was operating an insular grid with no international connections.

Worldwide, utilities were monopolies with little cost transparency. State authorities promoted nuclear energy by offering incentives. It has been demonstrated that not a single nuclear power plant has been built as a fully commercial project.1732 As of mid-2025, 60 units (or 95 percent) of the 63 reactors under construction in the world are being either built in nuclear weapon states or implemented by nuclear weapon state-controlled companies in other countries.1733 An important motivation for bearing the costs was to share efforts with military developments or to build up nuclear expertise that might be useful for military purposes (see Interdependencies Between Civil and Military Nuclear Infrastructures in WNISR2018 and Civil-Military Cross-Financing in the U.K. Nuclear Sector in WNISR2024).

Civil nuclear power was developed in the old paradigm of energy systems, and it has not substantially developed further since. Can it, nevertheless, provide answers to new energy problems?

Help in Solving the Climate Emergency?

A central argument in favor of the admittedly costly nuclear power is the ever-increasing urgency for decarbonization. As discussed in the opening section on the evolution of the overall context of nuclear power development, the speed at which fossil fuels need to be substituted is far beyond nuclear power’s achievable growth rate (see also Climate Change and Nuclear Power in WNISR2019). Growth rates are not only limited by the costs involved, but more so by the lead times necessary for building a single plant and for building up the necessary industrial capacities, workforce skills, as well as oversight and enforcement capacities that can ensure a level of quality, safety, and security throughout the entire fuel chain that is not required in other industries. Most scenarios aiming for decarbonization foresee a declining share of nuclear in the overall energy supply.1734

Nuclear power’s much longer lead times and far higher costs per MWh, compared to modern renewables, mean new nuclear plants will save less emissions per dollar as well as per year (see Climate Change and Nuclear Power in WNISR2019).

The argument that nuclear power would have a positive economic impact because of its need for a highly skilled workforce is not as straightforward as it seems. The nuclear industry requires highly specialized nuclear experts who need long training periods; highly skilled manual workers who can reliably execute tasks at changing, large, and remote construction sites and follow stringent safety standards in specialized industries; as well as specialists who can build unusually large constructions onsite. Such a workforce is much more difficult to build up and to coordinate than the equally skilled workforce needed for conceiving, building, and operating solar panel, battery, or electronics factories.1735

Bearable Nuclear Costs with Serial Production and New Designs?

The WNISR has extensively discussed the rising costs of nuclear power over the years. Earlier in this chapter, they have been explained from the perspective of fundamental scientific principles. There is no reason to think that the cost situation could substantially improve. New concepts cannot eliminate the basic cost drivers: safety and security issues linked to the elementary nuclear forces involved and the inherent drive for size in thermal power plants, whatever primary energy they operate on.

With some success, advocates of nuclear energy have spread the impression that decarbonization is expensive anyway and that alternatives to nuclear are no better. However, the widespread perception that the energy transition is an economic burden and that renewable energies are expensive has proven to be wrong. A series of simulations has consistently shown that 100-percent renewable energy systems are not more expensive than the conventional fossil-fuel-based approach.1736 More importantly, 100-percent renewables-based systems are less costly than those that include nuclear power. New technological and manufacturing advances presented in this chapter even indicate a larger cost difference.

Moreover, nuclear power plants with investment costs of some US$15–20 billion and a ten-year construction time represent a considerable cluster risk even for larger economies. In an increasingly dynamic environment, the importance of this risk is growing. Competing technologies today evolve rapidly; their unit size is much smaller, and their lead time is much shorter. Shorter product cycles mean faster innovation: Wind turbines built twenty or thirty years ago are often being “repowered”— substituted with much more efficient technologies—which frequently doubles or triples their output. In contrast, “uprating” nuclear power plants on the occasion of a lifetime extension can only increase yields by little more than 20 percent at the most. Repowering to double their output becomes attractive for solar power plants as well.1737 Most private companies remain reluctant to take the additional risk of nuclear power plant investments. Increasingly, as upfront subsidies and state responsibility for accident risks and long-term waste management were not sufficient to convince reluctant private operators, guaranteed offtake prices were introduced to also cover commercial risks at least during much of the lifetime of the plants.1738

A real problem for fast decarbonization with renewables is the cost structure involving high upfront investments. However, nuclear has no advantage in this regard either. The upfront capital costs of nuclear power plants far exceed those of renewable energies, leaving wide margins for storage and other flexibility investments. The capacity-factor-corrected CAPEX for nuclear power plants is about 3.5 times higher than for large PV plants.1739 And even marginal costs from old reactors are higher than the LCOE of solar.

The cost problem that led to a sharp decline in nuclear power deployment more than forty years ago has not disappeared; it has worsened.

Nuclear as Reliable Source for “Firm Power”?

Advocates of nuclear power often propose this energy source as an ideal complement for renewable energy that would be able to reduce the need for flexibility.1740 They promote nuclear power as an essential component to provide “firm power”, often keeping this concept rather diffuse, suggesting that nuclear power can provide electricity around the clock and throughout the whole year.

The concept of firm power has only emerged recently. Often, it is confounded with the concept of baseload. More precisely, it may be defined as follows:

Firm power generation represents the capability for a resource or an ensemble of resources to meet electrical demand 24x365, i.e., embodying both baseload and dispatchable generation capabilities.1741

Essentially, it mirrors the old paradigm that generation-plus-storage must be able to meet variable demand by providing a combination of baseload (nuclear, lignite, run-of-river hydro) and dispatchable (coal, gas, reservoir-based hydro) power generation. A slightly different interpretation of firm power is included in the effort to calculate integration costs of renewable energy (as discussed earlier).1742

Taking into account new decentralized storage technology, which can be deployed on the generation as well as on the demand side, this concept can be widened to describe a hybrid combination of generation and storage—similar to the concept of “virtual power plants”—able to optimize the meeting ground between flexible demand and flexible supply in such a way that energy flows are as continuous as possible for enabling maximum use of grid capacity.

Proposing nuclear power as a component in a firm power energy mix with strongly fluctuating renewable power, however, neglects its weak points. For technical and economic reasons, nuclear plants do not provide the type of flexible, dispatchable power that can fill the gaps between solar power peaks. They need flexibility from other sources for bridging considerable planned and unplanned outages and for buffering between changing demand and their inflexible full-load operation. The cost of their baseload electricity is more expensive than the ultra-flexible combination of renewables-plus-storage-plus-flexible demand.

The latest driver for nuclear power plants is the assumed rapidly increasing demand for energy supply in vast quantities for large AI data centers (“firm power for AI”).1743 It may be interesting for data-center operators to secure the large output of existing nuclear power plants for the large consumption of their facilities that changes within seconds, minutes, and hours, but not over weeks or months. However, they require guaranteed complementary grid power during the reactor downtime (around one month per year or more, plus any unplanned outages) and storage for managing the rapidly changing power demand in daily operations. Where utilities are eager to attract a new large customer, they may be willing to sign corresponding contracts.1744 These agreements, often incentivized by production tax credits in the U.S., de facto divert low-emission power from other customers and indirectly subsidize electricity supply to Big Tech companies (see United States Focus).

Building new nuclear plants for AI data centers, however, appears incoherent. Time horizons do not match: while data centers need power quickly, nuclear power plants need many years to build; competing solar power plants can be set up within months.

In several years, needs and priorities may completely shift: On the one hand, power needs may dwindle—conventional data centers have become much more efficient over short periods, so may specialized AI data centers, due to both software and hardware innovations. On the other hand, location preferences may change—data can be transmitted easily, and new centers may emerge where energy is cheapest.

The idea of building power plants especially for large industrial complexes with a high need for a rather continuous supply of power is failing for the same reasons. Nuclear plants need flexibility from the grid and powerful grid connections to compensate for their inflexibility and outages. There is no “firm power” advantage compared to the alternative set of technologies; cost, safety, security, and time horizon disadvantages remain. In other words, since nuclear electricity is uncompetitive, changing its use does not make it competitive, because cheaper carbon-free sources (or savings) can still beat its price.

In larger markets, other flexibility sources may compensate for the rigidity of a smaller share of nuclear power, but this comes at a cost. In smaller electricity markets though, such as in many low-income countries, substantial shares of nuclear power from a few plants can become a serious problem in the long term because they block other developments, but even in the short term in the case of unplanned outages of what would be a substantial share of capacity in the grid.

Remaining Drivers for Nuclear Energy Expansion

Against the background of all the facts and considerations in this chapter, one may wonder why nuclear power still has influential advocates. The initial enthusiasm for nuclear power and the surge of new nuclear projects ebbed more than forty years ago. Since then, with the exception of few countries, the traditional community of nuclear professionals has considerably shrunk. A handful of large incumbent international nuclear builders—only about a handful in the world have built abroad in recent years—have restarted hiring, hoping for a “nuclear revival”. Moreover, there is a new wave of heavily subsidized startups, especially around the development of SMRs, that attract increasing amounts of private capital. However, overall private investments in nuclear, both into innovation and into deployment, are marginal compared to those flowing into competitors of nuclear energy.1745

Under the pressure of the urgency of decarbonization, hard choices have to be made. New renewable and associated technologies have the potential to address the decarbonization issue. However, they also present major challenges to the established energy system paradigm, which has persisted for decades despite the decline of nuclear power. Western industries were reluctant to embrace these new technology developments, which are now dominated by Chinese industries. Those who are still attached to the old paradigm, whether due to vested interests, ideological reasons, or deep identification with past activities and declining structures, prefer to strive to renew the past glory of nuclear power than to embrace this strange new mix of technologies, which involves all levels of stakeholders including citizens who were never interested in electricity issues. Changing paradigms is a tedious and slow process. “Science advances one funeral at a time,” wrote Max Planck, who was at the center of the paradigm shift in physics a century ago.1746 But the time left for complete decarbonization of the energy system is short—not even a generation.

One of the reasons for the ongoing and even growing interest in nuclear energy points back to its origins. “Without civil nuclear, no military nuclear; without military nuclear, no civil nuclear,” said French President Macron in his historic speech on nuclear power in 2020.1747 Military considerations can override economic ones, and mostly, they are not explicitly stated as Macron did. When analyzing national nuclear power programs, the interdependencies with military interests are often obvious—the WNISR has shown this repeatedly. Maintaining a workforce with nuclear expertise, developing small reactors for submarines and aircraft carriers or simply showing the ability to carry out nuclear projects are essential to incumbent military nuclear establishments (see Civil-Military Cross-Financing in the U.K. Nuclear Sector in WNISR2024).

There are other powerful drivers like geopolitical interests and strategies. A nuclear power plant project has a fabulous lock-in effect that binds two or more countries together over decades. The discussion of these issues would far exceed the scope of this chapter.

Nuclear Power vs. Renewable Energy Deployment

Introduction: Global Trends

A mix of contradictory trends has marked the year 2024. On the one hand, in most regions, inflation, rising interest rates, growing political uncertainties, regulatory stagnation, and declining investor confidence provided a more difficult environment for the energy transition than the year before. On the other hand, declining prices and advancements in key technologies have been encouraging and have opened new perspectives. In the first months of 2025, these trends became even more pronounced. The energy transition is entering a new phase, and this is reflected in the structure of this part of the report.

The complexity of the energy system’s transformation, driven by the need for decarbonization and new technological opportunities, poses a challenge to traditional comparisons. As the share of renewable electric power generation in the system increases, nuclear power can no longer be simply compared with wind and solar power, which have quite distinct characteristics and behave differently within the system. Looking at the overall energy system—including the electrification of the heat and transport sectors, as well as the evolving needs and self-generation capabilities of electricity consumers—considerable changes to the potential systemic roles of renewable energies and nuclear power need to be examined. Therefore, this year, the WNISR includes a special focus chapter on the systemic aspects of the power sector (see Challenges of Integrating Nuclear Power into the Energy System). And the present chapter, in the tradition of past editions, examines the dimensions of deploying new energy technologies and includes new considerations, foremost among them the impact of more-affordable batteries. In this way, the aspects discussed in the short chapter on Power Firming and Competitive Pressure on Nuclear in WNISR2024 are integrated into larger contexts.

Batteries are becoming increasingly important. In an increasing number of electricity markets, the growing share of wind and solar-photovoltaic power, whose output depends on fluctuating natural phenomena, requires much more flexibility in the system—flexible consumption, storage, or complementary flexible generation. The cost of this flexibility may be challenging to calculate since it may be associated with changes in consumer habits, industrial procedures, or more sophisticated management of upgraded grids, and it may take time to implement such flexibility. More straightforward to calculate is the cost of electricity storage in batteries, and this has declined considerably in the course of 2024 alone (see Challenges of Integrating Nuclear Power into the Energy System). Think tank EMBER titled a report released in June 2025: “Solar electricity every hour of every day is here and it changes everything” and further stated “the LCOE of near 24/365 solar generation are falling fast – 22% in the last year alone.”1748

1. Solar Power Overtakes Nuclear Power

Source: EMBER, Monthly Electricity Data, July 2025

The full consequences of this will take some time to develop and to be understood. Key quantitative data in 2024 underline the growing importance of photovoltaics (PV):

• Compared to 2023, PV capacity additions in 2024 have grown by 33 percent (increasing the cumulative global capacity by 36 percent), according to SolarPower Europe. That is impressive, but still a smaller market growth rate than the “exceptional 85 percent surge” in 2023. Not only has a more difficult overall economic environment and a shifting political focus from climate concerns to security issues contributed to this outcome, but also declining electricity prices following the energy crisis in 2022/23.1749 Solar electricity generation has increased by about 28 percent.1750 In April 2025, global solar electricity generation exceeded nuclear power generation for the first time and kept doing so in May and June 2025 (see Figure 60).
• Wind energy deployment has been hindered by the general economic and political environment more than PV deployment, as technological and cost developments were less favorable. While investments declined more significantly, so did costs, so capacity additions were only one percent lower than in the previous year. Wind power electricity generation in 2024 still increased by close to 8 percent.1751
• Battery Energy Storage Systems (BESS) are characterized by two metrics: power capacity (kW or GW) and energy capacity (kWh or GWh). Globally, grid-scale BESS power grew by 113 percent in 2024 (reaching 126 GW). The average annual growth rate during 2014–2024 was 75 percent.1752 BESS energy capacity (expressed in GWh) has grown even more, as more and more batteries with longer storage time are installed.1753
• Between nuclear reactor startups and closures in 2024, nuclear added 5.3 GW net. Taking into account the balance of reactors entering and leaving LTO (see Figure 67) operating capacity increased by 2 percent (+6.9 GW) and electricity output by 2.9 percent.1754

Including hydro, gross electricity generation of all renewables grew by 862 TWh or 9.6 percent in 2024—not enough to keep pace with the rising global electricity demand (up by 1,293 TWh), driven by higher temperatures, industrial expansion, electrification, and unambitious demand-side management and efficiency policies.1755

In national and international politics, climate and energy policies are losing priority. Several countries are scaling back policies supporting the energy transition. This trend has increased globally since the Trump administration took office in the U.S. in early 2025. At the same time, renewable energies are becoming increasingly competitive, and their deployment dynamics are showing remarkable resilience considering the headwinds. The 2002–2024 constancy of exponential growth in global renewable generation suggested that market forces tend to beat political fluctuations, and early signs in 2025 suggest this may continue to be true.

This leads to the question of whether we have reached a point where the energy transition driven by economic considerations alone, without significant policy support, could halt global warming. In April 2025, Bloomberg New Energy Finance (BloombergNEF or BNEF) published a new, updated base-case scenario, which explores “how far and how fast the low-carbon transition can proceed based purely on competitive economics and existing short-term policy settings.” BloombergNEF concludes that under these assumptions, global warming would amount to 2.6°C in 2100, and that “deploying only economically competitive climate solutions will not be enough to avert climate disaster.”1756 Nuclear power generation, in this scenario, would increase around 20 percent between 2023 and 2035—certainly unrealistic considering nuclear’s long lead-times—while PV growth would amount to 450 percent.

Other voices are more optimistic concerning the economic push for renewables.1757 And this chapter will unveil some signs of a more fundamental shift.

China’s control of around four-fifths of every step in the global supply chain for batteries and photovoltaic modules has been met with increasing concern, primarily in the U.S. and Europe. However, the effects of the Biden administration’s efforts to establish a strong renewable energy industry in the U.S., which began to become visible in the investment statistics in 2023, have slightly receded in 2024. The Trump administration has rolled back many of the incentives for renewables and strongly expanded those for nuclear power and fossil fuels. The difficulties of substituting fossil and nuclear fuel imports from Russia, following the invasion of Ukraine, have continued to draw attention to the significance of energy independence and energy supply security on a global scale, but on these important issues too the advantage goes to renewable supplies and efficient use. Even if one buys solar modules from China, once they are bought, they require no further material inputs to keep working for decades.

Investment

Statistics describing the development of electricity generation include three consecutive types of indicators, each representing a distinct step on the path to the goal. Each of these indicators—investment sums (US$ per year), generation capacities (GW added per year), and electricity generated (TWh per year)—comes with its own interpretation difficulties. Investments are an early indicator of the system’s evolution. They translate into building generation capacities. Investing in the right technology is crucial—relative costs, technological and political risks, construction times, and cost development trends can make a big difference in the selection. Investment decisions can determine infrastructure structures for many decades, depending on the technology. Generation capacities, then, reflect past investment decisions and determine future electricity generation. Gigawatt indications per technology are not directly comparable as are dollars and euros, since full-load hours differ between technologies, technology generations, and locations. Terawatt-hours measuring the electricity generated, finally, are easier to compare, but, as we will see later, their value depends on the regularity or rhythm with which they are produced.

For two decades, investments in renewable power generation globally have exceeded those in nuclear energy—which hardly changed over the past decade—and are now 21 times higher (see Figure 61).

According to BloombergNEF data, investments in PV saw the strongest growth (22 percent) in 2024 compared to 2023. That is a rather good result, considering the growth rate of about 20 percent one year earlier, the growing inflation, and the shift in public attention from climate concerns to security issues (though those would support the same investments).1758 As a consequence, with some delay, taking into account the ongoing decline in PV system prices1759 (see next section), one can expect the corresponding PV generation capacities to continue to grow strongly.

Investments in wind power, on the other hand, have declined by 16 percent in 2024. The Global Wind Energy Council notes: “It might be fair to say that many people in the wind industry will not look back on 2024 too fondly. It was a year in which the impact of interest rate increases, inflation, supply chain pressures, investor confidence, regulatory inertia and political uncertainty all had a relevant impact across many key markets.”1760

1. Global Investment in Renewables and Nuclear Power, 2004–2024

Source: BloombergNEF, 2025

Notes:

As of the 2025 edition, WNISR is using BloombergNEF data as source for nuclear as well as renewables investments and is thus abandoning its own calculations.

The BloombergNEF nuclear investment data series is only available from 2015 onwards.

BloombergNEF notes: “Nuclear power investment data covers reactors under construction or major refurbishment. It excludes funding for research and development and capex investment during refueling outages. We also exclude spending to bring back retired reactors to operation, such as in Japan, and at Palisades and Three Mile lsland in the US.” Whereas WNISR’s nuclear investment numbers were based on the total estimated investment sum and allocated to the construction-start year, BNEF has opted “to spread the investment cost of each project equally over the construction period”. As shown in WNISR2024, when cumulated over several years, BloombergNEF and WNISR analyses were in the same order of magnitude.

Global investment into nuclear power projects remains small—below US$35 billion per year (including major refurbishment), in the 5-percent range of renewable energy investment. It has not seen any significant growth rate over the past decade, and only the Middle East and North Africa (MENA) region has seen a rather consistent growth. On the contrary, total nuclear investment has been shrinking slightly over the past two years. A closer regional breakdown of nuclear investments (see Figure 62) reveals ups and downs in investment patterns. The year 2024 shows a remarkable decline in the Americas (–26 percent) compensated by an expansion in China (+25 percent), and a decline in Europe (–10 percent) compensated by an increase in Russia (+60 percent).

1. Regional Breakdown of Nuclear Power Plant Investments, 2015–2024

Source: BloombergNEF, 2025

Notes: see under Figure 61.

The 2024 regional breakdown of renewable energy investments reveals interesting changes in dynamics compared to 2023 (see Figure 63 and Table 18). Despite the Biden administration’s efforts, the U.S. has slightly scaled back, while China has increased its efforts but has not yet reached its 2022 numbers, following the setback in 2023.

1. Regional Breakdown of Renewable Investments, 2015–2024

Source: BloombergNEF, 2025

Notes: Europe includes the EU27, the U.K., and the rest of EMEA (Europe, Middle East, Africa) as presented in Table 18.

Most remarkable, however, is the 2024-surge in minor markets: Asia-Pacific, excluding China and India, +81 percent; Central and Southern Africa, plus Europe excluding the E.U. and U.K., +85 percent; and the Americas excluding the U.S. and Brazil, +167 percent. Together, in 2024, they have seen an average 95-percent increase, while investments in the rest of the world have declined by 5 percent. Their share in global investments has increased from 13 to 23 percent (see Figure 61).1761 Persisting discrepancies between regions will be discussed below in the context of electricity generation statistics.

1. Regional Breakdown of Renewable Energy Investments (in US$ billion) and their Growth Rates, 2022–2024

Source: BloombergNEF, 2025

Notes: EMEA: Europe, Middle East, Africa; MENA: Middle East + North Africa.

Electricity Generation Cost Development

Lazard’s annual Levelized Cost of Energy (LCOE) analysis provides the most-quoted analysis of electricity costs by different technologies, although it primarily relies on U.S. data. As Figure 64 shows, solar and wind deliver far more competitive electricity than gas and coal, not to speak of nuclear, whose costs are three times higher. The newest edition, released in mid-2025, shows that electricity costs from utility solar have declined slightly (–4 percent), while onshore wind costs, though still highly competitive, have increased by around 22 percent.1762 Negative macroeconomic factors (interest rates, installation cost inflation etc.) had a strong influence, while the development of solar module manufacturing costs and prices, as well as wind turbine costs, proved more favorable.1763

1. The Declining Costs of Renewables vs. Traditional Power Sources

Source: Lazard Estimates, 2025

Notes: LCOE: Levelized Cost of Energy

*This graph reflects the average unsubsidized LCOE values in current dollars (not adjusted for inflation) for a given version of LCOE study. It primarily relates to the North American energy landscape but reflects broader, global cost developments.

While Lazard’s reporting is based on U.S. markets where import tariffs for Chinese manufacturers affected local prices, other sources based on other markets report an ongoing decline in the cost of PV. The German Fraunhofer ISE (Institute for Solar Energy Studies), the largest solar research organization in Europe, states in its latest Photovoltaics Report on global developments:

Due to the coronavirus crisis and the associated disruptions to supply and trade chains, market prices rose noticeably in 2022 and at times some products were not available in sufficient quantities. In 2023 prices fell again and have continued to fall in 2024.1764

The International Energy Agency’s Photovoltaic Power Systems Programme (IEA-PVPS) writes in its overview on global markets:

After several years of tension on material and transport costs, module prices continued to drop through 2024 in a still massively over-supplied market, putting tremendous financial pressure on all industrial actors on one hand but stimulating markets on the other.1765

The factors influencing LCOE calculations are illustrated in the most recent analysis by the United Nations’ International Renewable Energy Agency, IRENA (see Figure 65).1766 Total Installation Costs (TIC) for PV plants have declined everywhere but to differing extents. The capacity factor has evolved differently in different countries, as it also depends on the specific weather conditions in a particular year. The LCOE depends on the two previous inputs while also reflecting interest rates and other financing conditions.

1. Solar PV Cost and Performance Trajectories in Selected Countries, 2022–2024

Source: IRENA, 2025

These numbers, however, do not reflect the temporal availability of the electricity generated. This is particularly important for solar and wind energy, which depend on weather conditions. This conditioned availability leads to much confusion in the public debate, as an optimization must consider more factors than in the past. However, nuclear energy also requires compensation from other sources, as it needs to operate relatively constantly and requires extended downtimes for inspection, maintenance, upgrading, and refueling as well as unplanned outages for technical reasons.1767

Various authors have developed different methods to calculate “integration costs” or “firming costs” for intermittent/variable/fluctuating1768 energy sources that are not dispatchable. They are usually not taken into account for technologies that are deemed dispatchable (e.g., nuclear power) but have their own level of intermittency. The problem is that such costs depend on the specific conditions of the grid in which these sources are to be integrated—specifically, the temporal structure of electricity demand (which can vary widely due to building characteristics, peak-shifting technologies or behaviors, etc.) and the costs of flexibility required from other sources/sinks. Moreover, the yield of a solar or wind plant depends on its specific location. Therefore, measures such as the VALCOE—where VA stands for Value Adjusted—of the IEA vary across different countries and change over time (see also Challenges of Integrating Nuclear Power into the Energy System).

1. Sharp Decline in Battery Costs After the End of the Lithium Shortage

Source: Goldman Sachs, 20251769

The storage (or equivalent flexibility) required by a specific generation profile to deliver a constant output for a defined fraction of the year is a more useful benchmark, as it does not depend on the specifics of the grid. The think tank EMBER, in its previously quoted report on the impact of declining storage prices, employs such an approach. Pragmatically, EMBER starts from a very sunny location in the U.S. (Las Vegas, Nevada) and calculates that a 1 MW constant output over 97 percent of the time can be achieved with a combination of 5 MWp (megawatt-peak) PV and 17 MWh of storage capacity.1770

This PV/storage combination results from a cost optimization that yields electricity costs of US$104/MWh (US$0.10/kWh)—22 percent less than a year earlier, primarily due to lower storage costs. The cost is also well above that of a 10-MW ultrareliable all-solar microgrid producing 24/365 power at below US$80/MWh, built in four months in early 2025 by Redwood Energy in Sparks, near Reno, Nevada. That plant simply laid solar modules on the ground, rather than mounting them on metal racks set in concrete, then added 2–48 hours of storage by ~800 second-life electric-vehicle batteries. Redwood Energy is now engineering solar microgrids an order of magnitude larger. Their round-the-clock solar power is cheaper and more reliable than grid power in early 2025.1771

Calculations for other cities show the effect of less sunshine: Madrid can cover 88 percent of the time at US$114/MWh, Washington and Buenos Aires 81 percent at US$124, and Birmingham still 62 percent at US$160/MWh.1772 EMBER concludes, and Redwood Energy’s project confirms empirically:

The emergence of 24-hour solar generation marks a fundamental shift in how solar fits into the broader power system. With the ability to deliver electricity around-the-clock, solar can now support 24/7 clean energy contracts (PPAs) for industries which require continuous power, not just daytime supply. This is extremely valuable for emerging economies, where solar-powered industrial and economic zones can emerge in sunny areas far from existing grid infrastructure.

Essential for this result is the rapid decrease in battery costs. The disruptive decline in 2024 (40 percent) represents a return to the longer-term price curve, after the lithium supply chain finally adapted to the rapid growth in demand that had led to temporary shortages and price increases (see Figure 66).1773 A particularly striking example was the case of “’Mind-blowing bids in China Power’s 16 GWh tender” in late 2024. Reportedly, the “tender attracted 76 bidders with quoted prices ranging from US$60–82 per kWh, averaging US$66.3 per kWh.”1774 A staggering decline indeed if compared with Goldman Sachs’ estimate for average global storage costs in excess of US$100/kWh for 2024 (see Figure 66). And then in early June 2025, a 25 GWh auction “viewed as a watershed moment for the marketization of China’s energy storage industry” saw four-hour storage bids in the range of US$51–68/kWh averaging US$59/kWh, that is 11 percent below the December 2024 auction average.1775

With upcoming new technologies and increasing production—still mainly driven by the growth of electric vehicle markets—battery costs are expected to further decrease for many years to come. As in the past, the battery learning curve will most probably be steeper than that of photovoltaics, resulting in an accelerated cost decline of round-the-clock solar power. This will affect not only utility-scale installations but also small-scale installations behind the meter (see also Challenges of Integrating Nuclear Power into the Energy System).

Installed Capacity and Electricity Generation

The Overall Picture

In 2024, installed global electricity generation capacities increased by a slightly smaller percentage than the record increase of the year before, while nuclear power showed a 5–10 times lower growth rate than its renewable competitors:

• Solar power capacity: +32 percent (2023: +32 percent, 2014–2024: +26 percent/year);
• Wind power capacity: +11 percent (2023: +13 percent, 2014–2024: +12 percent/year)
  onshore +12 percent (2023: +13 percent), offshore +7 percent (2023: +20 percent);
• Operating nuclear power capacity: +2 percent (2023: +0.4 percent; 2014–2024: 1.1 percent/year);
• Grid scale battery capacity: +113 percent (2023: +121 percent, 2014–2024: +75 percent).1776

Generation capacities of these different technologies cannot be simply compared or added because their load factors (percentage of the time they are running on full capacity) are different. One must therefore look at the electricity generated for understanding the proportions (although generation lags behind capacity installation). Figure 67 shows the long-term dynamics of both: the different growth curves reflect the exponential growth of wind and solar—with wind starting first and solar catching up with a higher growth rate—in contrast to nuclear, which has essentially remained constant for much more than this quarter of a century. In 2025, both wind and solar are likely to catch up with nuclear output.

1. Global Wind, Solar, and Nuclear Installed Capacity and Electricity Production

Sources: WNISR with IAEA-PRIS, IRENA, Energy Institute, 2025

Notes pertaining to figures comparing production and capacities from various power sources. Unless otherwise indicated:

- Production data for renewables and nuclear are in net TWh, from Energy Institute “Statistical Review of World Energy 2025 – Consolidated Dataset”, June 2025;

- Gross production numbers from Energy Institute are used for comparisons with fossil fuels (for which net data is not available);

- Shares in total generation are based on gross data;

- Numbers for installed capacity for renewables are from IRENA; and

- Nuclear operating capacity (i.e., excluding reactors in LTO or Long-term Outage) and installed capacity (including reactors in LTO) are compiled by WNISR based on IAEA-PRIS.

Over the past decade, non-hydro renewables—mainly wind, solar, and biomass—have increased power generation by 236 percent while nuclear output gained just 9.4 percent (see Figure 68).

1. Nuclear vs. Non-Hydro Renewable Electricity Production in the World

Source: Energy Institute, 2025

Notes: see under Figure 67.

Looking at the evolution of the various sources’ relative shares in world electricity generation since 2000, shows that coal’s part peaked in 2007 at 41 percent and declined to 34 percent in 2024, while the renewables’ share (including hydro) increased from 19.4 percent to 31.6 percent between 2010 and 2024 alone. Over the same period, the nuclear share fell from almost 13 percent to 9 percent (see Figure 69).

1. Electricity Generation in the World by Source, 2000–2024

Source: Energy Institute, 2025

Notes: see under Figure 67.

In 2024, global gross electricity generation evolved as follows:1777

• Solar power generation: +28 percent, or + 461 TWh
  (reaching 2,112 TWh or 6.8 percent of total generation)
• Wind power generation: + 8 percent, or + 188 TWh
• Non-hydro renewable power: +14 percent, or + 670 TWh
  (reaching 5,415 TWh or 17 percent of total generation)
• Hydro power generation: + 5 percent, or + 192 TWh
• Nuclear power generation: + 3 percent, or + 80 TWh
• Fossil fuel power generation: + 2 percent, or + 334 TWh
• Total global electricity generation: + 4 percent, or +1,293 TWh (reaching 31,256 TWh)

This means that the additional non-hydro renewable electricity generation in absolute terms (670 TWh) covered only 52 percent of the overall worldwide increase in electricity production (1,293 TWh). As global electricity consumption has begun to grow more rapidly in recent years, this is even less than the share of 57 percent achieved over the past decade (2014–2024) during which non-hydro renewables increased annual production by over 4,000 TWh of a total net increase of just under 7,000 TWh, reflected in Figure 70.

1. Power Generation in the World, Annual Production Compared to 2014

Source: Energy Institute, 2025

In the rather unlikely case that the growth rates of 2024 remain constant, renewables would catch up with overall generation increase in three to four years, and solar power would cover total generation (equal to consumption, which would have nearly doubled) by 2038. However, most scenarios do not foresee such ongoing renewables growth due to political obstacles and challenges in adapting the whole energy system to large shares of fluctuating renewables. In fact, the whole system logic is undergoing a deep transformation, as the chapter on Challenges of Integrating Nuclear Power into the Energy System of this report describes in more detail. BNEF writes in its already quoted new baseline scenario:1778

Our base case scenario falls short of the Paris Agreement goals, with emissions falling only 22% by 2050 – a far cry from net zero – and is consistent with 2.6°C of global warming by 2100.

IRENA and the IEA keep emphasizing that present policies are not sufficient for reaching the pledge to triple renewable energy deployment and to keep climate change within 1.5°C.1779

Photovoltaics and Batteries—The New Dream Team

SolarPower Europe (SPE), in its Global Market Outlook published in May 2025, forecasts a decline in solar power capacity growth rates: After added capacity grew 33 percent in 2024, SPE’s medium scenario envisages only 10 percent in 2025, 1.5 percent in 2026, and resumes stronger growth with 14 percent in 2027. The main reason SPE cites for this temporary drop is a change in Chinese deployment policies. “With China implementing major changes to its solar market design this year, a temporary dip in global growth in 2026 appears very likely.”

However, recently, there have been several indications that China is not slowing down at all. January through May 2025, China had already installed 70 percent of the capacity added in 2024.1780 Moreover, in early July, the Chinese State Grid Energy Research Institute declared that for 2025, it expected an additional solar capacity of 380 GW, 36 percent up from last year.1781 Reported Chinese installations of 93 GW of solar and 26 GW of wind power capacity just in the month of May 2025 are equivalent to nearly four gigawatts per day.1782

Predictions of a slowdown would fit the history of the systematic underestimation of photovoltaic development. Not only does the IEA have a long record of heavily lowballing photovoltaics, but even SolarPower Europe (SPE), the association of the European PV sector, systematically undervalued its own technology: Its 2020 Global Market Outlook forecasted less than 200 GW of new capacity for 2024. Every year, this figure was increased until the latest Outlook stated that, effectively, 596 GW had been installed in 2024, three times SPE’s 2020 estimate for 2024.

Widespread expectations that the expansion of renewables and especially solar will slow down are rooted in grids’ difficulties to provide sufficient flexibility for high shares of renewables. Figure 71 shows the increase in hours with negative power prices in Europe. Negative power prices indicate a temporary overproduction, which cannot be avoided at all times. However, if they last for too many hours, they are an indicator of the system’s difficulties in coping with fluctuating renewables, and they reduce the income of solar installations in particular. Lawmakers and grid operators in various markets have answered very differently to these challenges. Those reluctant to embrace change call for a slowdown in the growth of renewables and are slow to pave the way for additional flexibility. Others adapt market and connection rules to allow potential flexibility sources, such as storage, to participate in electricity markets and in other ways encourage grids to catch up with the fixed-price, nearly-zero-operating-cost renewable supplies that need to be matched with customers. Batteries offer lucrative opportunities to buy negative-cost or cheap renewable power, sell it on-peak at high prices, and meanwhile sell valuable ancillary services such as frequency and voltage stabilization.

1. Growth of Hours with Negative Power Prices

Source: Rystad Energy, “Energy Storage Outlook—Energy Macro Report”, May 2025

Notes: Rystad Energy defines Europe as the EU27 + the U.K. and Ireland + the EFTA countries (Norway, Iceland, Switzerland, and Liechtenstein). The number of negative hours has been calculated based on data from all the power markets in continental Europe. Thus, the cumulated number is larger than the average 8,766 hours per year.

With more temporal distance, it may appear that 2024 was the year when the energy system crossed a tipping point.

The rapid decline in battery prices (see Figure 66) may be a game changer that will remove hurdles for a sustained rapid growth of PV. While batteries are being mentioned here and there in recent scenarios and outlooks, the impact of recent price drops has not yet been integrated. With more temporal distance, it may appear that 2024 was the year when the energy system crossed a tipping point.

Starting at negligible levels five years ago, the deployment of stationary Battery Energy Storage Systems (BESS) has increased significantly since then (see Figure 72).

The acceleration of grid-scale BESS installations is impressive. Over 80 percent of the overall BESS capacity has been added in the past three years. This overall picture, however, conceals a much more complex dynamic across different markets: Large grid-scale batteries presently make up around four-fifths of the worldwide market, while small batteries behind the meter—appealing to other customers and having a different impact on the system—are one-fifth.1783 Both the overall adoption of batteries and the proportion of those in front and behind the meter differ strongly between regions.

1. Grid-Connected Battery Storage Additions, 2020–2024

Sources: BloombergNEF, based on Graphic by PowerSwitch/Volta Foundation, 2025

Considering only grid-scale BESS power, between 2018 and 2021, the average annual global growth rate was 52 percent and then jumped to 115 percent (2021–2024). The main driver was the Asia-Pacific region, whose growth rate jumped from 51 percent to 167 percent, resulting in a 68 percent share of global installations by 2024. In the EU27, growth only climbed from 44 percent to 59 percent annually, reaching a share of 3.8 percent. The U.K. alone accounts for more than all EU27 installations.1784

U.S. data for 2023 show how tightly the deployment of BESS is linked to photovoltaics. Of the 19 GWh grid BESS energy capacity added in 2023 (+166 percent), 79 percent were paired with PV plants, one percent with other power plants, and 20 percent were not paired. Interestingly, 89 percent were owned by independent power producers, which can then better compete with incumbent utilities’ legacy thermal power plants.1785

While globally, the overall storage capacity of around 360 GWh1786 at the end of 2024 makes up for only about 6 percent of the average daily solar power production, in some regions, batteries are already starting to have a considerable impact. The public grid in California, with the assistance of BESS, has successfully shifted solar electricity generated around noon into the evening hours.1787

To put present growth rates into perspective: extrapolating the 2024 growth rates of PV electricity generation (+28 percent) and BESS energy capacity (+87 percent), batteries would be able to store the average daily solar electricity production in 2032, a quantity that would correspond to nearly half of the present daily overall electricity consumption.1788 This does not predict the future but it does illustrate the enormity of current growth rates.

Extrapolating the 2024 growth rates of PV electricity generation (+28 percent) and BESS energy capacity (+87 percent), batteries would be able to store the average daily solar electricity production in 2032.

These astonishing prospects may even prove conservative, because they include only stationary batteries. Yet a fifth of new cars in the U.S. and a third in the world are battery-electric, rapidly expanding what will soon be an immense resource—about 350 GWh of batteries-on-wheels in the U.S. Many new electric vehicles (cars, trucks, buses, etc.) are designed for smart power exchanges with the grid, charging when power is cheap and abundant, then selling back electricity (and other services) when that is profitable and the vehicle is parked. Many new business models and institutional structures are rapidly emerging to exploit this potential. There may soon be a flood of batteries that could reshape grid operations and economics before a new reactor could even be built. And meanwhile, similar forces are at work with stationary storage in buildings, which use half the world’s and three-fourths of U.S.’ electricity.1789

First Signs of a Boom Behind the Meter

While cost-effective utility-scale storage may remove barriers for further growth of PV in many electricity markets, small-scale storage behind the meter may become even more important on a global scale.

The role of prosumers, who consume a portion if not all of the electricity they produce without passing through the public grid, has been a topic of discussions for many years.1790 Recent massive cost reductions for batteries and energy management systems have been game-changers. They enable much higher rates of self-consumption, making prosumers more independent of public policies governing feed-in tariffs and of grid companies that are late to adapt to renewable energies. There are first signs that this could lead to a surge in PV, even where governments and incumbent utilities are reluctant to embrace new renewable technologies.

The focus of international reporting on PV has always been on utility-scale installations. However, recently, rooftop installations and their potential impacts are increasingly gaining attention. For installers and component providers, rooftop installations have simply addressed a different kind of customer. However, on the one hand, with a growing share in the electricity market, fluctuating renewables are becoming increasingly challenging to integrate, requiring adjustments in regulation and grid infrastructure. On the other hand, with increasingly affordable storage, electricity generation and consumption behind the meter are becoming more independent of the public grid and will thereby alter and diminish its role.1791 The significance of this change has not been fully acknowledged yet. The last IEA report on the subject was published in April 2024, specifying behind-the-meter storage prices for 2025 that amount to twice the actual market prices in July 2025.1792

The last IEA report on the subject was published in April 2024, specifying behind-the-meter storage prices for 2025 that amount to twice the actual market prices in July 2025.

Over the past decade, the global market share of rooftops, as reported by SolarPower Europe, has varied annually between 28 percent (in 2016/17) and nearly 50 percent (in 2022).1793 Concerning additional installations in 2024, SolarPower Europe reports a minor reduction of the global rooftop market share from 43 percent to 42 percent, due to a further drop in electricity prices following the energy crisis in 2022–2023 and a series of policy changes that reduced financial support. However, among the Top 10 markets, Brazil (place 4) with 63 percent rooftops and Türkiye (place 7) with 90 percent show the rooftop potential in sunny countries, while Germany (place 5) with 62 percent and Japan (place 9) with 68 percent deliver the proof of rooftop’s attractiveness for highly industrialized, less sunny regions. Some of these figures should rise further as, for example, the Tokyo Metropolitan Government intensifies its push for ubiquitous rooftop solar for greater Tokyo’s 35 million residents, e.g., through the mandatory installation of solar power systems in residential buildings starting in April 2025.1794 China, which had pushed the rooftop market share to 58 percent in 2022, scaled back to 44 percent—still above the international average, which is suffering from a mere 17 percent rooftop market share in the U.S. and 15 percent in India.1795 .

Cumulative PV installations behind the meter never seemed to be of interest, but with the prospect of massively raising the self-consumption rate with the help of batteries, they start to get more attention. In 2024, the IEA’s PV observatory PVPS published a list of the rooftop share of cumulative installations for 27 countries in 2023: Globally, 44 percent of all installations were “decentralized”. Outstanding high percentages were found in Türkiye (79 percent), Germany (78 percent), Italy and South Africa (both 69 percent), as well as Japan (60 percent).1796

Combining new PV installations with storage and retrofitting existing ones could have a tangible impact on public grids. In Germany, already in 2023, 80 percent of new rooftop PV installations were equipped with batteries.1797 A new phenomenon in Germany is “balcony power plants”. Over a million such small, self-installed PV plants make up only one percent of all PV capacity, but their ongoing boom, following a doubling in 2024, shows the attractiveness of easy electricity production behind the meter, even without compensation for electricity fed into the grid.1798 Most of them are now coupled with batteries to increase the self-consumption rate.1799 Many need no installer—you just plug them into the wall outlet to feed their modest amount of power back into the home’s circuits whenever available.

However, all these figures need to be taken with a grain of salt. Statistics concerning rooftop installations and, more so, concerning electricity generation and self-consumption are not equally reliable in all countries. And the statistics of international organizations sometimes differ. A recent report from IEA-PVPS provides a detailed explanation of the difficulties.1800 A prominent example of misinterpretations is the case of Pakistan (referred to in Challenges of Integrating Nuclear Power into the Energy System), where a mysterious drop in power consumption despite economic growth was later explained by analyzing import statistics: private, non-registered PV installations had allowed for considerable self-consumption that did not pass through the public grid. With growing shares of photovoltaics behind the meter, statistics that reflect only what happens on the public grid increasingly reveal only one part of the story, making forecasts more difficult. This is now important in Australia, where PV production is on most homes and counts as a negative load.

Moreover, it must be acknowledged that all these markets are still in a very early stage, where huge price spreads often lead to confusion and obscure opportunities. A case in point is the cost of residential PV plus battery installations in the U.S., which are still three times higher than in Australia1801 or two times higher than in Germany. No wonder that the rooftop share in the U.S. (17 percent) is much lower than in Australia (67 percent), the country with the highest PV density in the world (1,521 W/capita). Also in northern Europe, where the cloudy Netherlands has the second-highest PV density worldwide (1,491 W/capita), residential PV is much cheaper than in the U.S.

For residential and commercial users, the question of energy supply is getting more and more complex: The first step was to complement electricity from the plug with an own PV system, which would sell most or all of the electricity to the grid and involve complicated tariffs. Then batteries appeared on the horizon, soon accompanied by Home Energy Management Systems (HEMS) for optimizing not only the long-discussed time of use of the washing machine but also those of more demanding devices such as the heat pump, the air conditioning, and the wallbox for charging the electric car. This requires a learning process not only for customers, but also for the vendors of all these devices.

The theoretical potential given by today’s battery and PV costs is far from being exhausted. The main topic at trade fairs for new energy technologies, such as the Intersolar in Munich, is management systems, their interoperability and standardization, their transparency for the customer, and their rapid evolution with the support of AI.1802 Even large equipment suppliers are still in the early stages of their journey to provide convincing solutions, as evidenced by the offers of leading strategy consultants.1803 With more user-friendly solutions, managing the complexity of integrated renewable energy systems at the building level becomes rapidly more attractive.

Now add the vehicle-to-grid opportunity mentioned above. Several European firms sell energy and valuable ancillary services back to the grid. One, the Swiss-based Mobility House, has identified 21 such offerings. A few years ago, selling 13 of those earned an average of about €1,000 (US$1,100) per car per year. In Ireland, where high renewable output makes real-time electricity prices fluctuate, one operator estimates potential revenues many times that large. Some automakers (or other aggregators) may dispatch customers’ collective vehicle batteries as a major grid resource, compensating owners for any diminished battery life and ensuring no compromise in driving capability or experience. The financial benefit to owners could be large. And if rolled into the vehicle’s purchase price, the effective discount could sell more electric vehicles sooner.1804

Persisting Large Gaps Between Regions

Major international organizations such as the IEA, IRENA, World Bank, and REN21 have identified the gap in renewables growth between low-income and other countries as a major issue. IRENA states:

Asia has kept its leading position since the past few years, accounting for 71% of new renewables capacity in 2024, followed by Europe and North America (respectively contributed 12.3% and 7.8% to the addition), leaving a huge gap with Africa, Eurasia, Central America and the Caribbean which together only accounted for 2.8% of total renewables capacity addition. Despite its massive economic and development opportunities, Africa only increased its renewables capacity by 7.2%.1805

As capacities do not allow for direct comparison between technologies, electricity generation is more revealing. Table 19 shows the contribution of solar, wind, and nuclear to overall electricity generation in relation to the world average.1806

1. Solar, Wind, and Nuclear: Regional Electricity Generation Metrics vs. World Average in percentage

	North America	S&C America	Middle East	Africa	EU27	Asia-Pacific	World	U.S.	China	Australia	Germany	Saudi Arabia	UAE
Contribution to total electricity generation compared to world average (in percent)
Solar	90	107	44	32	160	112	100	98	123	264	221	27	127
Wind	116	120	4	40	231	87	100	124	118	152	384	5	1
Nuclear	182	19	29	9	258	56	100	197	50	-	-	-	254
Per capita electricity generation compared to world average (in percent)
Solar	215	97	86	5	259	97	100	342	226	727	336	94	282
Wind	278	112	8	6	351	78	100	430	226	398	529	15	3
Nuclear	434	17	56	2	416	48	100	688	91	-	-	-	565
Total	239	90	195	17	161	87	100	349	184	276	152	354	222

Sources: Electricity generation data from Energy Institute; Population data from Our World in Data; and WNISR calculations, 2025

Note: Continents and regions correspond to traditional geographical grouping used by the Energy Institute.

Africa shows a particularly low deployment of solar and wind power. When considering the generation per capita, the differences are even more extreme, as per capita electricity consumption in the E.U. is ten times higher and in North America 14 times higher than in Africa on average. This is alarming as population growth in the coming decades will likely mainly occur in Africa. Natural conditions on this continent, on the other hand, are ideal for developing PV, as solar irradiation is higher and seasonal differences are smaller than in places farther from the equator. Just as cellphones leapfrogged over landline phones, solar power is an increasingly powerful tool for development, especially in rural and peri-urban areas where grid power is absent, slow, and costly, but local entrepreneurs and simple finance (even via scratchcards) are rapidly exploiting that gap.

For many years, distributed renewables have been proposed to provide electricity-based energy services to rural communities, but volumes have remained low.1807 The 2024 surge in investment in minor markets highlighted earlier in this chapter may indicate an improvement. However, the recent cuts to foreign aid by the Trump administration are causing a reshuffling of energy infrastructure financing across the continent, which hurts many programs and projects.1808 Some argue this could also be an opportunity for greater reliance on local resources.1809 New calculations using recent battery costs reveal very interesting amortization times for small-scale projects also in urban areas.1810 After the withdrawal of the U.S. and aid reductions by Europe, China is said to become an increasingly attractive funding partner for large-scale projects. This may involve increasing interest in nuclear energy.1811

Another region, in Table 19, showing little progress with renewables is the Middle East—despite excellent conditions for solar energy and no lack of regional capital. The share of solar in total electricity production in the region is less than half of the worldwide average. Of particular interest are Saudi Arabia and the United Arab Emirates.

Saudi Arabia, the world’s top petroleum-exporting country, achieves only 37 percent of the solar energy per capita compared to the E.U. and one-quarter of the total electricity generation share compared to the global average. However, recent efforts to build a post-petroleum economy are showing some results with 163 percent growth in solar energy generation in 2024 vs. 2023.

The UAE, ranking eighth in petroleum production, and proudly hosting IRENA since its foundation in 2009, has achieved above-global values for the solar share in electricity generation, but still considerably lower than in the E.U., while wind energy is nearly absent. On the other hand, the UAE’s nuclear power share in overall electricity is far above the global average, exceeds that of the U.S., and is nearly on par with the E.U. (see Table 19).

With President Trump’s recent u-turn in American AI and chip policy, local electricity generation policies in the region gained additional international visibility. The U.S. has agreed to help the UAE and Saudi Arabia become AI powers and build large data centers.1812 AI geopolitics is at the center of the international security debate,1813 and data center electricity consumption is central to the discussion on the rapidly growing demand for highly concentrated firm dispatchable electricity.1814 As a key hub for U.S. efforts to maintain AI dominance, the region will see considerable growth in electricity consumption. Agreed U.S.-supported AI hubs in the UAE are designed to comprise data centers consuming 5 GW. Khazna, the infrastructure provider, is betting on solar. Nuclear and gas, however, are also in the still somewhat opaque power mix.1815

Status and Trends in China, the European Union, India, and the United States

China

China dominates the world of solar and wind power in a double sense: It dominates the value chains for producing solar and wind power equipment on the global markets, and it dominates the deployment of solar and wind power with outstanding numbers at home.1816

In 2024, China alone accounted for 40 percent of the global solar electricity generation, and the same is true for wind. But solar grew by 44 percent, more than three times as fast as wind, which grew 13 percent. In contrast, China’s nuclear power accounted for 16 percent of global nuclear electricity production and grew only by 3.7 percent. Growth rates of total electricity, solar, and wind power were all around 55 percent higher than at the global level, while nuclear lagged behind with a 27 percent higher than global growth rate. Compared with national electricity production, solar power grew 6.5 times as fast, wind 1.9 times, and nuclear 0.6 times—i.e., nuclear power lost market share.1817 On a per capita basis, as we saw earlier, Chinese solar power production is less impressive, reaching two-thirds of the level of the U.S. and Germany and only one-third that of Australia (see Table 19).

Figure 73 and Figure 74 show, in a historical perspective, that the evolution of renewable power in China has followed a coherent, continuous strategy. From 2020 onwards, the nuclear curves are flattening compared to wind and solar. After a modest increase in nuclear power, wind has taken the leading role, followed by solar. In 2024, solar and wind generated over four times more power than nuclear (see Figure 73). Since 2010, output of solar increased by a factor of over 800, wind by a factor of 20, and nuclear by a factor of six (see Figure 74).

1. Wind, Solar, and Nuclear Installed Capacity and Electricity Production in China

Sources: WNISR with IAEA-PRIS, IRENA, Energy Institute, 2025

Notes: see under Figure 67.

1. Nuclear vs. Non-Hydro Renewables in China, 2000–2024

Source: Energy Institute, 2025

Note: see Figure 67 for comparison.

As Figure 75 illustrates, compared to overall electricity generation, the share of nuclear power has even declined in recent years. After reaching its apex in 2021 with 4.8 percent—about half of the global value—the share of nuclear power dropped by 0.1 percentage points every year to 4.5 percent in 2024. It is also remarkable that the share of coal—while still excessively high—has declined by more than 20 percentage points, from over 80 percent in 2007 to less than 58 percent in 2024.

1. Electricity Generation Mix in China, 2000–2024

Source: Energy Institute, 2025

Notes: see under Figure 67.

For more than a decade now, China has made considerable efforts to also lead in the battery space: first by electrifying two-wheelers, then by deploying electric buses and cars, and not long ago by entering the stationary electricity storage business. In global battery supply chains, too, China controls over 80 percent of every step.1818 At the end of 2024, 60 percent of all grid-scale BESS were installed in China.1819 Indeed, in 2024, China increased new grid-connected battery capacity by 130 percent to reach a total of 74 GW/168 GWh.1820 Batteries will likely further boost the adoption of solar power.

In every phase of the energy transition, at least since the beginning of this century, China has been instrumental in cracking the resistance of incumbent industries in America and Europe by creating and dominating markets that offered highly competitive energy technologies. The outstanding average annual GDP growth rate of 12 percent between 2000 and 20241821 has given leeway for new industries, without too much opposition from domestic incumbents. Meanwhile, the renewables industry is a key part of a strong, independent economic force in China labelled “clean energy” that also comprises electricity grids, energy storage, EVs, railways, and nuclear power, and it contributed 10 percent of China’s GDP and a quarter of China’s GDP growth in 2024.1822

European Union

As Figure 76 illustrates, in comparison with the corresponding graphs for China, India, and the U.S. (see Figure 73, Figure 78, and Figure 79), renewable energies took off much earlier in the E.U. than in other continents. However, growth has not been as steep as in China.

1. Wind, Solar, and Nuclear Capacity and Electricity Production in the EU27

Sources: WNISR with Energy Institute, IRENA, and IAEA-PRIS, 2025

Notes: see under Figure 67.

In 2024, wind contributed 17 percent to the overall electricity generation, while solar contributed 11 percent. Nuclear power still outpaced both solar and wind individually with a 23 percent share, of which 59 percent was generated in one country, France. The wind plus solar contribution of 28 percent to the total electricity production, however, not only outcompetes the nuclear share in the E.U., but is also much higher than in China (18 percent) or the U.S. (17 percent) and nearly twice the global average (15 percent). Figure 77 shows how the growth of renewables is bringing down the consumption of fossil fuels, with solar and wind alone generating just about as much power as all fossil fuels combined.

European growth rates for solar and wind in 2024, however, have been underwhelming, with solar growth (20 percent) at 0.7 times the average global rate and wind growth (2 percent) at 0.2 times the global rate. In contrast, nuclear power has made more progress in Europe than globally, achieving a growth rate of 1.65 times the worldwide average and surpassing nuclear growth in both China and the U.S.—although, with only 4.8 percent, it is still not even a quarter of the E.U.’s solar growth.

In fact, the E.U.’s nuclear growth is exclusively due to production in France that saw recovery from some very bad years (see France Focus). French reactors increased generation by 41.3 TWh in 2024, while the overall EU27 output increased by only 28.4 TWh. In other words, in the rest of the E.U., nuclear output declined by 13 TWh.

1. Electricity Generation in the EU27 by Fuel, 2015–2024

Source: EMBER, 2025

However, these overall numbers for the EU27 (which accounts for 72 percent of the electricity generation of the continent without Russia) conceal considerable differences within Europe. The two largest E.U. economies, Germany and France, have followed very different strategies.

In 2024, nuclear power accounted for 67 percent of electricity generation in France. In Germany, it was zero. In Germany, wind and solar energy made up 43 percent, while in France they accounted for 13 percent. Solar electricity in Germany grew by 16 percent, while in France it grew by 10 percent. The growth rate of overall electricity generation in France was 9 percent with domestic consumption increasing by only 3 TWh or 0.7 percent. French electricity exports jumped 48 percent to 103 TWh for record net exports of 89 TWh.1823

An intensely used trading system interconnects European countries, making comparisons more difficult. On balance, in 2024, France exported 16 percent of the electricity it generated (the quivalent of around 130 percent of its solar plus wind or 23 percent of its nuclear power generation). Germany imported about 5 percent of its consumption, while solar plus wind provided 41 percent. Italy, also without its own nuclear power, imported 16 percent of its consumption, while generating 32 percent of its consumption with renewables.1824

Increasingly, the E.U. institutions influence the member states’ energy policies, although the European Commission is not directly responsible for energy policy as it is for trade policies. Presented in December 2019, the European Green Deal, encompassing a wide range of policies, aims at making Europe climate-neutral by 2050. It has provided a strong push for renewables across all E.U. countries. European institutions are urging for more interconnection capacities between countries and for other steps to facilitate a truly integrated European energy market. This would increase overall flexibility and encourage optimal siting of generation and consumption. Current energy generation statistics reflect past decisions and policies.

Meanwhile, the political landscape has changed. Since 2022, energy-supply-security considerations and geopolitical threats have increasingly influenced the agenda. For the political majorities in the E.U. and its member states, accelerating the energy transition has become less of a priority compared to a year ago. SolarPower Europe writes in its mid-year analysis of the E.U. renewables market:

After exceptional expansion in 2022 (+47%) and 2023 (+51%), growth flattened at 3.3% in 2024. For 2025, the market is expected to contract for the first time in nearly a decade, with a projected – 1.4% growth under the Medium Scenario. The current downturn is driven primarily by the rooftop segment, particularly residential solar systems. In many major Member States, primarily households and SMEs are postponing investment in solar installations following lowering electricity price trends and weakened support frameworks. In many cases, rooftop solar incentives have been withdrawn or scaled back without effective alternatives, resulting in a short rush and sudden market decline.

Looking at the economics of solar power, the objective conditions are not as bad as the sentiment, which has been influenced by the hefty anti-renewable policies of the Trump administration. Small private investors immediately react to changes in financial support before returning to a more rational approach, as is happening with the heat-pump and e-mobility markets in Germany.1825

In Europe, following Russia’s full-scale invasion of Ukraine and the subsequent energy crisis, energy-supply security is gaining importance, perhaps more so than on other continents. While climate concerns, therefore, might have lost some of their priority, renewables and the energy transition are increasingly being considered as essential tools for gaining energy-supply security.1826 Also, in a direct military sense, the security advantages of renewables that have been evident in Ukraine are being appreciated in the E.U.1827

India

As Figure 78 shows, renewable capacities have continued their growth path, while annual wind generation stagnated, caused not by missing capacity growth but by weaker-than-usual winds.

However, the PV capacity growth of 34 percent corresponds to the world average, while PV electricity generation grew by 17 percent, versus 28 percent globally. That means there is hope that India is catching up in the global context, as wind plus solar generation still accounts for only 11 percent of total electricity generation, compared to the 15 percent worldwide average. On a per capita basis, it looks much worse: Indian PV power generation reaches only 36 percent of the world average, wind only 18 percent.1828

Nuclear, on the other hand, has seen a 13 percent generation increase in 2024, which in absolute terms (+6.5 TWh) accounts for just over one third of the solar increase (+21 TWh).

1. Wind, Solar, and Nuclear Installed Capacity and Electricity Production in India

Sources: WNISR with Energy Institute, IRENA, IAEA-PRIS, 2025

Notes: see under Figure 67.

In its efforts to triple solar power by 2030, the Indian government bets strongly on expanding rooftop installations. The Ministry of New and Renewable Energy reports just under 19 GW “grid-connected solar rooftop” as of end of June 2025.1829 A net-metering program is supposed to boost adoption, but meets several hurdles, not least the opposition of utilities. It may well be that not all installations behind the meter—estimated at 5 GW—are appearing in the statistics.1830

United States

Unlike in China, the E.U., and India, and unlike the worldwide average, wind plus solar electricity production in the U.S. is still lower than the output of nuclear reactors. Nuclear power generation has been relatively stable over the past twenty years, but in 2024 it grew slightly, making up for losses in recent years.

Wind and solar in 2024 have grown respectively by 3.5 percent and 28 percent in capacity, and by 7.7 percent and 27 percent in electricity generation. That corresponds quite precisely to worldwide growth rates. U.S. wind and solar together produced 17 percent of total electricity generation, slightly exceeding the 15 percent worldwide. With 14.8 percent of global electricity generation, the U.S. is a heavyweight, but consumes less than half as much electricity as China.

1. Wind, Solar, and Nuclear Installed Capacity and Electricity Production in the United States

Sources: WNISR with Energy Institute, IRENA, IAEA-PRIS, 2025

Notes: see under Figure 67.

Within the U.S., there are significant differences between states. Unlike in Europe, there is no nation-wide electricity market. The rules of electricity markets differ considerably, and so do the shares of renewables and nuclear, as well as their developments.

Most notable is the development in Texas, traditionally a conservative oil state governed by Republicans, where traditional proponents of strong support schemes for renewables would not expect a boom in renewables. However, the deregulated energy-only market with nodal pricing, authorities that ensure no monopolies can cripple free competition, streamlined permitting processes, and a reluctance toward state subsidies have led to a surge in solar that has outpaced California in 2024. Between 2015 and 2020, utility-scale solar power capacity experienced an average annual growth rate of 73 percent, which somewhat slowed to a still remarkable 46 percent between 2020 and 2024.1831

In recent years, Texas has also become the epicenter of BESS growth. By the end of 2024, Texas had an installed capacity of almost 9.7 GW of grid-connected batteries, expected to more than double in 2025 to 20.1 GW. Only three years earlier, in 2022, battery capacity stood at 2.8 GW. At the same time, fossil-fueled plants have been retired, notably 7.3 GW of coal and gas closed between 2018 and 2023.1832

Under the Trump administration, however, this economically motivated success story risks slowing down. In a political climate that opposes renewables and favors fossil fuels, local politicians try to pass legislation to prioritize fossil fuels.1833 Texas legislators in early 2025 nearly passed a new law severely restricting and penalizing new and existing renewable generation but reversed course, apparently influenced by local renewable investors and beneficiaries and by fears that without fast-growing renewables, the state could not keep its lights on.1834

The struggle in Texas reflects an overall trend in the U.S. The Trump administration is attempting to reverse the efforts of the Biden administration to accelerate the energy transition through a wide range of initiatives. In 2024, the Biden efforts, especially with the so-called Inflation Reduction Act or IRA signed into law in 2022, were unfolding. But with the election of Trump and his first executive orders starting 20 January 2025, the wind changed. While many local Republican officials are defending the positive effects of the law on their local economies, essential institutions and programs are being severely damaged.1835 Many observers, however, are maintaining that the economic strength of renewables will limit harm by the new federal policies, as many local and state governments continue to support the transition, many corporations continue to buy increasing amounts of renewable power to get its low and stable prices, and markets tend to beat politics.1836 See also United States Focus for details.

Conclusion: A Pivotal Year

From this year’s analysis of the deployment of renewables and nuclear power, several key points emerge that corroborate the systemic considerations in the preceding chapter. These developments may later be considered to have been a turning point in the history of worldwide electricity and energy systems.

• Solar electricity generation is closing in on advancing wind energy and hardly-increasing nuclear power generation, and it continues to grow steeply while increasing its cost advantage. Photovoltaics has consolidated its role as the forthcoming central pillar of energy systems. In April 2025, for the first time, solar outpaced global nuclear power generation on a monthly basis, and it did so again in May and June 2025.
• The significant decline in battery storage costs has reached a tipping point. The deployment of stationary battery systems both at utility scale and behind the meter has started to soar in some geographies at unprecedented growth rates and is changing the economics of renewable power. Despite fluctuations in generation, solar and wind power are becoming available to deliver power around the clock at very competitive costs.
• Signs are amplifying of a revolution behind the meter, as seen by powerful examples in Pakistan and South Africa. Combining rooftop solar with batteries now allows for high degrees of self-sufficiency at increasingly competitive costs everywhere, and it is already winning cost comparisons in some regions. Building owners decide and invest on their own to improve energy supply security and drive down costs, including mobility. Avoiding or minimizing permitting and land-use issues, as well as excluding grid access problems—usually associated with utility-scale solar—considerably accelerates implementation.
• The global divide between rich and poor countries concerning renewable deployment may start to shrink. While actual differences in generation figures are still scandalous—especially considering the phenomenal potential for renewables in the south, solar above all—investment data show considerable improvement. Indeed, microgrids with PV and batteries offer attractive alternatives where public grids are weak. The global divide in PV equipment manufacturing, however, remains: China dominates global markets.
• Nuclear power, unlike renewables, is confronted with a declining market share. Capacity and production are stalling. Despite an intense campaign for the revival of nuclear energy, investments have increased only marginally in the global energy landscape. Governments favoring nuclear power are finding they can have only as much of it as they can force taxpayers to pay for, since capital markets generally perceive too much financial risk for too little return in a dauntingly competitive market already over 90 percent captured by renewables.
• Political priorities are shifting. While climate change contributes to an increasing sense of insecurity and anxiety about the future, its priority on the political agenda has declined. Security, including energy-supply security, and economic concerns have become more critical. In this political climate, important forces are seeking relief with “proven solutions” and teaming up with incumbent interests. Others emphasize that renewables provide the best energy supply security and are now the least-cost option. In this debate, the political and economic weight of the renewables sector, both in terms of capital and of employment, has continued to grow. The main obstacle to further acceleration is widespread misunderstanding and misrepresentation of renewable power’s reliability when properly integrated with demand- and supply-side resources to keep the grid in balance.

There are two countervailing trends. On the one hand, renewables have reached technical, economic, and systemic maturity. It’s no longer mainly the political will and the wish to mitigate climate change, but above all the sheer economic competitiveness of new technologies and systems that has started to vigorously transform the energy system. On the other hand, political forces tied to the old system and new geopolitical drivers are hindering the transformation. Policymakers must decide whether their goal is to enable the new energy system or protect the old one. Those choices have major consequences; some soon, some later.

While the political battle around energy policy is intensifying in several countries, economic forces are moving on. Under political pressure, especially from the U.S. government, companies have started “greenhushing”, as opposed to greenwashing. The magazine The Economist reports that “companies which are continuing to take action to decarbonise have grown more reluctant to parade their efforts.”1837 It is becoming increasingly challenging to predict the impact of politics and policies on actual developments, and fortune-telling is not within WNISR’s mandate. Renewables have become much more independent from subsidies and support frameworks. On the other hand, against the backdrop of China’s dominance in renewable equipment markets and rising geopolitical tensions, trade and industry policies have become more influential.

Independent reporting on factual developments, therefore, is increasingly important for understanding what happens in the energy and electricity sectors. Available statistical reports are insufficient. This chapter illustrates that differentiating between installations in front of and behind the meter as well as monitoring in detail the developments in storage and grid systems is indispensable. More generally, for understanding change in the global energy system, it is not sufficient to look at the supply side—it never was. Consumption structures and the interconnecting infrastructure and technologies are essential parts of the picture and are transforming rapidly.

Annex 1 – Overview by Region and Country

Africa

South Africa

South Africa hosts the only commercial nuclear power plant on the African continent, consisting of two 900-MW reactors located at Koeberg, near Cape Town. Both reactors started operating in the mid-1980s. In July 2024, the plant passed the end of its originally projected 40-year lifespan. Following extensive upgrading work, the operating license of Koeberg-1 has been renewed in July 2024 for an additional 20 years, while Koeberg-2 has been permitted to keep operating until November 2025, by when its owners, South Africa’s national power utility Eskom, hope to gain a license extension to run also this unit for another 20 years.1838

With the country having in recent years experienced a prolonged period of acute power shortages, the decision to keep the Koeberg plant operational for a further 20 years, while not uncontested,1839 always seemed likely because insufficient new power generation capacity was under development to compensate for Koeberg’s 1800 MW. With power shortfalls still looming and efforts to reduce carbon emissions in South Africa’s coal-dominated electricity generation,1840 nuclear newbuild—in addition to the projected growth in renewables—is being actively considered1841. While discussions regarding the optimal way forward continue, an updated roadmap defining the country’s preferred future electricity provision modes is expected to be finalized soon.1842

In 2023, South Africa had been experiencing persistent power cuts and record electricity shortfalls that at times exceeded 6 GW.1843 2024 saw an unexpected and remarkable turnaround, with rolling blackouts ending in March until the end of the year.1844 Some of the factors responsible for the improved outlook were detailed in WNISR2024. The good run continued into 2025, although rolling blackouts again had to be implemented over several short spells, typically of about three days, between 31 January and 25 April.1845 One of these, from 7–10 March, was explicitly in part blamed on an unplanned outage of the single unit then operating at Koeberg.1846

Developments Related to Koeberg’s Lifetime Extension

Once again, only one of Koeberg’s two units has been operational at a time for practically the entire reporting period. This was at first due to the need to complete the upgrading work at Unit 1 required to secure a 20-year lifetime extension beyond the plant’s projected 2024 closure date. Once that exercise had been accomplished for the second unit in December 2024 after a year-long outage,1847 this was soon followed by the renewed shutdown of Koeberg-1 for refueling, maintenance and tests to the containment structure expected to be completed by July 2025,1848 a procedure that is later to be replicated with Koeberg-2, where the next refueling and maintenance outage has provisionally been scheduled for 8 September 2025 to 27 March 20261849. Hence, Koeberg is expected to continue operating at maximally 50 percent capacity for most of the remainder of 2025.

To approve the 20-year lifetime extension of Koeberg, the South African National Nuclear Regulator required a series of maintenance operations and instrumental replacements to be carried out. The most significant of these was the replacement of the two sets of three steam generators. These replacements and associated work were finally completed at the end of December 2024,1850 more than two years after the originally projected end date.

An initial 7-month long attempt to replace the steam generators of Koeberg-2 in 2022 was ultimately aborted. Thereafter these replacements and related work were performed on Koeberg-1 from December 2022 to November 2023. Koeberg-2 was then again switched off on 11 December 2023 to undergo the same work (see WNISR2023 and 2024 for more details). The date for the return to service of Koeberg-2, which had in April 2024 been projected to be in September of that year,1851 was again not met. The unit was finally reconnected to the grid on 30 December 2024 with full output achieved ten days thereafter.1852 For approximately one month in early 2025, both Koeberg units operated together, the first time since 2022 that full capacity could be maintained for such a period.1853

The final costs of the upgrades and related activities linked to the Koeberg lifetime extension certification have not been made public to date. These had in 2010 been estimated at ZAR20 billion (US$20102.7 billion),1854 and in 2024 a senior source at Koeberg quoted a figure of ZAR25 billion,1855 but several observers believe these are likely to be significantly higher,1856 especially considering that the South African Rand-US Dollar exchange rate changed from about ZAR7.5 to US$ in 2010 to about ZAR18 to US$ in 2025.1857

In July 2024, the operating license for Koeberg-1 was renewed for 20 years to 21 July 2044.1858 The National Nuclear Regulator also granted a short-term extension to the existing operating license of Koeberg-2 to 9 November 2025, arguing that it only came into commercial operation in November 1985, more than a year after Koeberg-1 had reached that milestone.1859

Other developments related to the Koeberg lifetime extension include an outage of the only recently refurbished Koeberg-1 unit starting on 11 September 2024 that led to ten days at zero power output.1860 The initiating event was only made public five days after it happened, and was ascribed to a failure of one of the isolation/block valves.1861 The incident ironically occurred just five days after an operational safety assessment by the IAEA expressed satisfaction that apparently most of the recommended measures to deal with previously identified safety issues had been attended to.1862 Outstanding items of concern include the need to test the effectiveness of the containment structures, as this is one of the procedures planned during the current refueling and maintenance outage.1863 Early in 2025, the Southern African Faith Communities’ Environment Institute (SAFCEI) launched a, as of mid-2025, still ongoing legal appeal against the decision to grant Koeberg-1 the 20-year life extension, arguing that it should not have been issued until all safety issues previously identified had been adequately addressed.1864

In 2023, Koeberg’s power generation had dropped to 8.13 TWh, almost 20 percent down from the previous year. In 2024, the annual output dropped further to 7.8 TWh, with a corresponding share of only 3.8 percent in the national electricity mix.1865 On 2 March 2025, the refurbished Koeberg-2 was forced to shut down for a week due to a “steam leak on the reheat system”.1866 While the exact cause of the problem was not explained further, a mention was made of the discovery of two unspecified “defects” that necessitated a further delay in the restoration of operation.1867 This unplanned outage was also one of the contributors to the national electricity shortfall that precipitated one of the spells of rolling blackouts in 2025.1868

Developments Related to Potential Newbuild

In December 2023, it was announced that a Request for Proposals (RFP) for nuclear newbuild projects would be issued by March 2024.1869 The intention to initiate work on adding 2.5 GW of new nuclear capacity was gazetted in January 2024.1870 SAFCEI and Earthlife Africa, two organizations that had in 2017 led a successful legal challenge to stop the then envisaged 9.6-GW Russian-led nuclear newbuild,1871 again aimed to obtain a court order to halt the RPF on the basis that it had been issued without the stipulated level of consultation. Instead of waiting for this legal process to unfold, the Minister decided in August 2024 to withdraw the RFP and to initiate a more thorough consultative process.1872

In early January 2024, the Ministry of Mineral Resources and Energy had published a draft of a new Integrated Resource Plan for Electricity (IRP).1873 The IRP is effectively a roadmap that guides South African government interventions to develop electricity generating infrastructure, and is intended to be updated every couple of years. The document published in January 2024 was effectively withdrawn for redrafting after it received unusually severe and widespread criticism for a range of demonstrated flaws.1874

In December 2024 a revised draft IRP was distributed to selected stakeholders for consideration. This version includes a schedule presented as the “Proposed Balanced Plan” in which two new nuclear generating units of 1250 MW each would come on line in 2036 and 2037 respectively, with two further 1350-MW units becoming operational in 2038 and 2039 respectively.1875 This is presumably the scenario that was preferred by the developers of that iteration of the IRP, though whether subsequent discussions endorsed this scenario has not been made public.

Various analysts are still very critical of the revised draft IRP, e.g., stating that “Technologies such as gas and nuclear were effectively forced into the scenarios despite the model not selecting them as optimal solutions.”1876 In May 2025, the IRP was reported to be still under discussion at the National Economic Development and Labour Council (Nedlac), described by the Minister as the final step before he would table the IRP at the Cabinet.1877

Given past pronouncements by the Minister of Electricity,1878 it is likely that additional nuclear power generation will be included in the revised IRP. Once released, the new IRP will also indicate if the projected new nuclear would be in the form of traditional large-scale plants or SMRs, with interest groups representing both these options actively lobbying. However an inclusion in the IRP should merely be seen as a statement of intent, and many proposed developments listed in past IRPs have failed to materialize due to financial, technological, and/or other reasons. The South African Nuclear Energy Corporation (NECSA) reportedly provided the South African Electricity Minister with a ZAR60 billion (US$3.3 billion) cost estimate for a nuclear newbuild, an amount that would according to the Minister, if actually realizable, not be financeable through the national fiscus.1879

Past efforts to initiate nuclear newbuild had been seen to favor Russia’s Rosatom.1880 While Rosatom retains a presence in South Africa, there are now signs of increasingly significant other partnerships linked to energy projects. China in particular has become more directly involved in solar plant construction,1881 and this increased activity now also encompasses nuclear projects1882. An offer from Iran to assist in potential nuclear builds was widely reported,1883 but any such activity would not be substantial given that Iran has no full-scale nuclear plant provider capacity.

In the area of waste management, the National Radioactive Waste Disposal Institute (NRWDI) has indicated that it intends to take over the management of the Vaalputs radioactive waste disposal site from the current operator, NECSA.1884 As the national statutory body for radioactive waste management, the NRWDI is authorized to take this step, which it views as an income generating opportunity.

The development of small modular reactors has again been frequently talked about in South Africa. Most significantly, Electricity Minister Ramokgopa indicated that he favored restarting the Pebble Bed Modular Reactor (PBMR) program that had been actively driven in South Africa in the years up to 2009, when it was abandoned due to rising costs and lack of progress (see South Africa Focus in WNISR2023).1885

While the period July 2024–June 2025 saw much improved electricity supply and greatly reduced power cuts in South Africa, this was largely the result of increased generation from the country’s existing coal plants. In 2024, 86.2 percent of Eskom’s electricity was produced from coal (84.0 percent in 2023), an unchanged 3.2 percent was contributed by solar, while the wind and nuclear fractions slightly dropped from respectively 5.9 percent and 4.1 percent in 2023 to 5.4 percent and 3.9 percent in 2024.1886

These figures exclude electricity generated through “behind the meter” solar rooftop installations, which has not been properly quantified, but appears to have increased substantially at the height of the South African power crisis, highlighted by the fact that Eskom’s electricity distribution has decreased from 222 TWh in 2019 to 206 TWh in 2024.1887

One way to estimate trends in South African solar energy installations is to quantify solar panel imports. Given the dominance of Chinese-built solar appliances in South Africa, that figure may be determined from Chinese PV export data.1888 Such data shows a major acceleration of Chinese-built solar panels entering South Africa from equivalents of 1.3 GW in 2022 to 4.3 GW in 2023, which then only slightly dropped to 3.8 GW in 2024 (Eskom’s power supplies improved in parallel).1889 While the actual added solar capacity is subject to significant uncertainty, with for example 2023 the addition estimated to be a substantially lower 3.3 GW1890 elsewhere, the growing migration to alternatives to Eskom electricity is evidently now lowering Eskom’s future electricity sale projections and thereby weakening the case for further nuclear builds.

The Americas

Argentina

In 2024, Argentina’s three nuclear reactors,

• Atucha-1, start of operation 1974, net capacity 340 MWe,
• Atucha-2, start of operation 2016, net capacity 693 MWe, and
• Embalse, start of operation 1984, net capacity 608 MWe

produced 10.5 TWh of electricity (10 TWh net), a 6.9 percent share of overall electricity generation and an increase from 9 TWh in 2023.1891 At its highest in 1990–1991, nuclear power’s share was about 14.4 percent (even 19.8 percent according to PRIS data) with just 56 percent of the current nuclear capacity available, which indicates substantial growth of non-nuclear electricity generation in Argentina (see previous versions of WNISR). Atucha-1 was built by Siemens-KWU, while the construction of Atucha-2 was initiated by Siemens-KWU, then abandoned, and later completed by the operator Nucleoeléctrica Argentina S.A. (NA-SA) with the help of external contractors. Apart from a demonstration plant in Germany, the design of both Atucha reactors, a pressurized heavy water design with natural uranium fuel, heavy water cooled and moderated, was built only in Argentina. The CANDU (Canadian Deuterium Uranium) reactor at Embalse was built by Atomic Energy of Canada Limited (AECL).

Atucha-1

Atucha-1 was designed for the equivalent of 32 years of full-power operation, which was reached in 2018. Nucleoeléctrica aims for a lifetime extension of 20 years. Following a framework agreement with the regulator, the prolonged operation of the unit was divided into two phases, each conditioned by specific regulatory requirements. Stage A of the lifetime extension began in 2018. Since the unit had already reached the end of its original lifetime, continued operation required amendments to its operating license, allowing it to operate until September 2024. Stage B shall begin once the reactor resumes operation after undergoing extensive refurbishment, modernization, and inspection work. That second phase of long-term operation is expected to last from early 2027 to 2046; the refurbishment outage began in September 2024 and is scheduled to last about 30 months.1892

While the reactor core design of both Atucha units features lower neutron fluxes in the reactor pressure vessel walls, making them less susceptible to neutron embrittlement than other pressurized water reactors, the new design lifetime will be close to 70 years, even if the license is still conditional on passing the Periodic Safety Reviews (PSR) every ten years. Retrofitting will certainly help, but reaching current state of the art safety standards will be very difficult if not impossible.

Atucha-2

Atucha-2 was ordered from Siemens-KWU in 1979 as a larger version of Atucha-1. As a consequence of the 1982 Falklands War, the construction was delayed because Siemens-KWU lost access to licensed technology for Argentina and was forced to develop substitute technology. Construction came to a complete halt in the 1990s following a major economic crisis in Argentina. By the time Argentina was ready to restart construction works, Siemens-KWU ceased to exist as a company and Framatome, the successor, had no intention of finishing the construction. Nucleoeléctrica decided to complete the construction on its own. Finally, grid connection occurred on 27 June 2014, but it took until 26 May 2016 to reach commercial operation.1893

According to IAEA-PRIS, the load factors of both Atucha units remained mediocre in 2024 at 68.8 percent and 67.6 percent, respectively. For information on the operational history of the Atucha plant so far, refer to previous editions of WNISR.

Embalse

Embalse, which started operating in 1983, was shut down at the end of 2015 for major overhaul, including replacing hundreds of pressure tubes, to enable it to operate for up to 30 more years, to 2049. Like Atucha-1, to continue operation, in addition to extensive work linked to the lifetime extension program, Embalse must perform a major PSR every ten years. It eventually returned to service in May 2019, with the refurbishment project estimated to have cost US$2.15 billion.1894 In August 2019, the regulator (ARN) renewed the operating license for ten years to 2029, when the next PSR is due.1895

Other nuclear activities: CAREM25, ACR-300, and Atucha-3

While 2024 was a rather smooth year for Argentina’s operating power reactors, major decisions were taken about the country’s long-term projects, CAREM25 and Atucha-3.

Already in 1980, the development of an Argentinean PWR, CAREM (Central Argentina de Elementos Modulares), with only 25 MWe began.1896 The work was first performed jointly by engineering and manufacturing company INVAP (Investigación Aplicada) and the National Atomic Energy Commission (Comisión Nacional de Energía Atómica or CNEA) but was then taken over by CNEA alone. The construction of a demonstration plant began near the Atucha site in February 2014, with startup planned for 2018. In 2005, CNEA had estimated construction cost at US$105 million,1897 but by construction start in 2014, estimates had risen to ARS3.5 billion (US$2014433 million).1898 In 2021, estimated cost had further increased to US$750 million,1899 and by 2024, over US$600 million had reportedly been spent with a further US$260–300 million needed for completion.1900 By then, commissioning was expected in 2028, a decade later than envisaged at construction start.1901 Following a change in government after the election of Javier Milei as president of Argentina in 2023, the chairperson of CNEA was replaced. The new CNEA president, Germán Guido Lavalle, declared the end of the CAREM project in late 2024. CAREM would not fit into the “commercially viable” category, as he reportedly said in an internal statement “let’s be intellectually honest, we are not going to sell 50 CAREMs, we know that…this reactor is not economically competitive. You only have to stand in front of the construction site to realize that this is not a small modular reactor.”1902 As reported in WNISR2024, this would have been the most expensive nuclear power plant per MW ever built, and construction had been halted due to budget cuts imposed under Milei which prompted a “critical design review”.

However, another project might take over as the Argentinean SMR flagship project. In 2024, INVAP received a patent from the U.S. Patent Office for the ACR-300 design.1903 The patent covers a small, compact pressurized water reactor with steam generators and pressurizer directly attached to the reactor pressure vessel, with three or four loops. According to some sources, an unspecified U.S. investor had been found to finance the design’s refinement to a 300 MWe three loop reactor, and funding for a design company with roughly 50 scientific employees was allocated.1904

While the current government of Argentina will not pursue efforts of recent years for a Chinese-built Hualong reactor at Atucha-3 for geopolitical reasons1905 (for the history of the Atucha-3 project refer to previous editions of the WNISR), the ACR-300 might take this position in planning, including up to four units at Atucha.

In late 2024, President Milei, flanked by IAEA Director General Rafael Grossi, announced the creation of a national Nuclear Plan and an Argentina Nuclear Council in charge of developing it.1906 The head of the newly formed council, Demian Reidel, reportedly stated:

The first step in this plan is the construction of an SMR at the Atucha site. Although the technology is new, thanks to the advanced state of engineering, technical support from the IAEA, and the firm political commitment of President Milei, there is a high probability that Argentina will be the first nation to produce and commercialize this innovative nuclear reactor model. This milestone will not only ensure our energy sovereignty, but also allow us to replicate this success throughout the country and export this technology to the world.1907

In May 2025, Reidel explained the first phase to be that of operating four SMRs with a total capacity of 1.2 GW within five years.1908

Power Mix and Energy Policies

According to the Energy Institute, in 2024, Argentina’s electricity generation was dominated by natural gas (54 percent), followed by hydro (19.5 percent), non-hydro renewables (15 percent), nuclear (7 percent), oil (4 percent), and other sources including coal for the remaining share.1909 There were only small changes in the electricity mix from 2023 to 2024.

Regarding new nuclear projects, Argentina demonstrated that due to long planning and construction times of nuclear power plants the risk of sunken investment in case of political change is considerable. While previous Argentinean governments were committed to the CAREM project, the current government, which is still pro-nuclear but with an emphasis on economics, abandoned 40 years of investments, only to foster a completely new design, which might need another two decades to arrive at a demonstration plant. Only time will tell if the political environment offers 20 years of stable support.

Brazil

Brazil’s two commercial nuclear reactors—Angra-1 and -2—are operated by state-controlled company Eletronuclear at the Central Nuclear Almirante Alvaro Alberto (CNAAA) site and in 2024 provided 15 TWh (net), an increase of 8.8 percent over the previous year, or 2.1 percent (a negligeable increase of less than 0.1 percentage points) of national electricity. Construction of the third reactor, that resumed at the end of 2022, was halted again in 2023, and is considered “suspended” in WNISR’s data.

The first contract for Angra-1 was awarded to Westinghouse in 1970. The 609-MW PWR went critical in 1981 and received a license for a 20-year lifetime extension in November 2024, allowing for operations until December 2044.1910 Eletronuclear estimates that total investment for lifetime extension works amounts to BRL3.2 billion (US$497 million), of which BRL2.5 billion (US$389 million) are to be spent between 2025 and 2029.1911

Angra-2 is a German-designed PWR with a capacity of 1275 MW that was connected to the grid in July 2000, 24 years after construction initially started. A 30-year license set to expire in 2041 was issued in 2011, but Eletronuclear has announced that it will likely request a 20-year extension.1912

The Angra-3 saga is one of the most chaotic in modern energy history in Brazil, the largest Latin American economy and the eleventh in the world. Preparatory work for the 1340-MW PWR designed by Siemens/KWU started in the early 1980s. It is unclear how much progress was made before a lengthy interruption starting in 1986. Construction eventually restarted in June 2010—the unit was then to be commissioned in 2016—but has been interrupted or stalled several times since. The administration of then-President Jair Bolsonaro (2019–2022) picked up efforts to complete the project setting in motion a major shakeup of Brazil’s energy industry landscape. However, the succeeding government of President Luiz Inácio “Lula” da Silva was ultimately left with the task of issuing the formal decision and setting the conditions for the decades-old project to be completed. Beyond the lengthy assessment phase and difficult decision-making process at the federal-level, Eletronuclear met (and seemingly solved) new legal disputes in the past two years, first with local government which withheld the necessary permits for resumption of civil works, then with the consortium hired to implement them, leading to the termination of the contract in June 2024. (See WNISR2023 and WNISR2024 for an overview of past events.)

In September 2024, Brazil’s National Bank for Economic and Social Development (BNDES) submitted a critical and long-awaited study to Eletronuclear and the Ministry of Mines and Energy. This financial assessment is legally to serve as basis for the National Energy Policy Council (Conselho Nacional de Política Energética or CNPE) to reach a definitive decision and determine the financial model of the project. The study, which had been commissioned in 2019, confirmed that about BRL12 billion (US$20242.2 billion) had already been invested in the project since 2010, and that completion would require a further BRL23 billion (US$20244.3 billion), while abandoning the project would cost BRL21 billion (US$20243.9 billion). The proposed tariff amounted to BRL653.31/MWh (~US$2024121/MWh), with commercial operation expected in 2031.1913 Per latest announcements construction is about two thirds complete.

“New inefficiencies or other delays in the work can no longer be incorporated into the price to be approved, making the respective costs a risk for the concessionaire, and no longer costs to be covered by consumers.”

The Minister of Mines and Energy, who chairs the CNPE—composed of other ministers and appointed councils—grants the project his unwavering support. However, upon convening on the matter since receiving BNDES’ assessment, the council postponed its decision twice,1914 and as of mid-2025, a new date had not been set.

Apparently weary of the possibility of new delays and cost-overruns, and well-aware of tariff raises that came with each renewed attempt at resuming the project, the Federal Court of Accounts (TCU) consistently reissues the recommendation that “new inefficiencies or other delays in the work can no longer be incorporated into the price to be approved, making the respective costs a risk for the concessionaire, and no longer costs to be covered by consumers.”1915

However, a further development that could significantly impact the terms of the financial model arose when in February 2025, Eletrobras and the federal government reached a settlement agreement—in a broader power struggle over state influence in the company’s governance—exempting Eletrobras from future investment in the project.1916 Eletrobras is Brazil’s biggest utility and, until its privatization in 2022, it was the parent entity of Eletronuclear. In order to maintain nuclear activities under state control, major restructuring was performed at the time, resulting in the creation of Empresa Brasileira de Participações em Energia Nuclear e Binacional (ENBPar) which took over Eletronuclear. Eletrobras however maintained a 67.95 percent interest in Eletronuclear’s capital, and a minority voting share of 35.9 percent. In 2022, ENBPar and Eletrobras entered into an agreement which kept the latter on-the-hook for financing the completion of Angra-3. Under the new agreement with the federal government, signed in March 2025, Eletrobras is to provide BRL2.4 billion (US$424 million) for Angra-1’s lifetime extension, while the 2022 Investment Agreement is suspended and is to be terminated “if and when” resumption of construction at Angra-3 is decided.1917

Nevertheless, Eletrobras remains “liable for any existing guarantees [it] provided for Eletronuclear’s financings prior to [its] privatization,” which as of December 2024, amounted to BRL6.1 billion (US$20241.1 billion) according to the company. It is now looking to sell its shares warning that

Even if the Settlement Agreement becomes effective, there are factors beyond our control that may affect its success. For example, BNDES may not be able to design a structure that ensures the economic and financial viability of the Angra 3 project that will be suitable for, and approved by, all stakeholders.1918

In the absence of CNPE’s approval, Eletronuclear President Raul Lycurgo confirmed at a parliamentary hearing on 27 May 2025 that the Angra-3 project remains frozen. Lycurgo reminded that Eletronuclear spends close to BRL1 billion (US$176.5 million) per year to maintain the project alive, including BRL100 million (US$17.6 million) on staff wages, BRL120 million (US$21.2 million) for mothballing equipment, and BRL800 million (US$141 million) for servicing debt on project financing.1919 A heavy burden for Eletronuclear which has already implemented various austerity measures, including cutting departments and management positions.1920 The company also requested a deferral of its debt payments to Caixa and BNDES until 2026, which had previously been suspended from July to December 2024.1921 In March 2025, Minister Silveira penned a letter asking his fellow Minister of Finance to intervene for the two banks to grant such a waiver, pointing to ENBPar communication warning of the “imminent risk of insolvency” of Eletronuclear and “the urgent need for actions to ensure the financial sustainability of Eletronuclear, given that, despite the efforts already undertaken by the Company’s management, such as a significant reduction in operating expenses and extraordinary measures to generate resources, the cash flow forecast for 2025 shows the persistence of the financial deficit, jeopardizing the operational continuity of the Angra 1 and Angra 2 facilities.”1922

As a TCU representative said in May:

The cost of not making a decision about what to do with Angra 3 is the worst of all. The Brazilian nuclear program depends heavily on Angra 3.1923

A thought seemingly shared by President Lula, who in June 2025, reportedly insisted for the CNEP to retable the issue and come to an agreement by end of the year.1924 If and when the project is greenlit, a program has to be developed, and a tender organized before an Engineering, Procurement and Construction (EPC) contract can be established. Experience shows such a process can take years.

Despite the uncertainty that Angra-3 casts over the industry’s mid-term prospects, further newbuild initiatives are already being sporadically discussed (see WNISR2024), including the possibility of deploying Small Modular Reactors (SMRs).1925 The National Energy Plan 2050 (PNE2050), released under the previous administration, had included the technology as potential component of Brazil’s energy mix starting “after 2030”. More recently, cooperation with Russia on SMR deployment has reportedly been high on the agenda.1926 The country has also embarked in the development of its own 3–5-MW “microreactor”, which it hopes to bring to commercialization by mid-2033 to 2035.1927

However, the latest approved Ten-Year Energy Expansion Plan, released in April 2025, sees no capacity additions from nuclear in the entire reporting period up to 2034, except from Angra-3 which is mentioned as still subject to studies and approval.1928 SMRs are described as promising, but still having major challenges to overcome, such as technological or regulatory.

Under the Bolsonaro Government a new regulatory agency, referred to as ANSN (Autoridade Nacional de Segurança Nuclear) was reassigned CNEN’s (Comissão Nacional de Energia Nuclear) responsibilities to monitor, regulate, and inspect nuclear activities and facilities. CNEN remains in charge of planning, overall policy, and advocacy for nuclear energy.1929

But since its inception in 2021–2022, ANSN has not functioned (see also past WNISR editions) due to lack of staffing at all levels. The Court of Accounts alerted in March 2025 about the lack of a proper management plan, an appropriate budget, and adequate personnel. For that matter, it gave 120 days to the government to deliver a sound management plan to guarantee a proper functioning of the agency.1930 As of mid-2025, there is no indication that such a plan has been submitted.

As reported in previous editions, Brazil is expanding its uranium enrichment capacities and expects to service the entire fuel supply requirements of its then three (potential) reactors by 2037.

Energy Trends

In 2024, renewables (including hydro) represented 87.3 percent of power generation. Hydroelectricity, the main source of power, reached 413.2 TWh (gross) in 2024, while output from wind increased by 13 percent to 108.5 TWh and from solar by close to 41 percent to 71.3 TWh. Their respective share in the power mix stood at 55.4 percent (hydro), 14.5 percent (wind), and 9.6 percent (solar). Brazil is the world’s fourth largest producer of wind power and the sixth largest of solar PV.1931

According to IRENA, Brazil added 3.8 GW of wind capacity (+13 percent) and an impressive 15 GW of solar (+40 percent) over 2024.1932

However, both the output and share of fossil fuels increased in 2024, representing a combined 10.1 percent—compared to 8.6 percent the previous year—with oil at 1.4 percent, coal at 2.2 percent, and gas at 6.5 percent.

In November 2024, Brazil, who will host COP30 in November 2025, announced its new target to reduce greenhouse gas emissions in 2035 by 67 percent compared to 2005-levels.1933 The government is also working on updating its National Climate Plan ahead of COP30.1934 The country is trying to regain its climate leadership after the effects of Bolsonaro’s regime. Though prompt to implement such a reversal, the current administration has since passed sets of contradictory policies, including on oil drilling. With the global focus on COP30, the international attention will also be directed towards Brazil’s domestic policies and measures to achieve its targets, including the decarbonization of its energy sector. It is now clear, that nuclear expansion will not play any role, at least not in the short term.

Canada

Canada’s reactor fleet consists of 17 CANDU reactors with a total net capacity of 12.7 GW. Two reactors (Pickering-1 and Pickering-4) were closed on 1 October 2024 and 31 December 2024 respectively.1935 As of June 2025, three reactors (Bruce-3, Bruce-4, and Darlington-4) are being refurbished since 28 February 2023, 31 January 2025, and 18 July 2023 respectively.1936 Two of the units being refurbished have reached Long-Term Outage (LTO) status as of mid-2025.

According to Statistics Canada, nuclear reactors supplied 81.74 TWh (net) in 2024, slightly lower than the 84.57 TWh produced in 2023.1937 In 2024, nuclear energy constituted 13.4 percent of the total electricity generated in Canada, a slight decline from 2023. All but one of the nuclear reactors are located in the province of Ontario, where nuclear power contributed 50.9 percent of the electricity generated in 2024, below the 53.2 percent in 2023.1938

Refurbishment

Canada is in the process of refurbishing several of its ageing CANDU reactors, which “involves replacing core reactor components” such as “fuel channels, feeder pipes, calandria tubes and end fittings.”1939 The reactor that was most recently refurbished was Darlington-1, which took close to two years and nine months for this process to be completed in late 2024.1940 According to the Darlington Refurbishment Program Annual Report from December 2024, the program had cost CAD11.1 billion (US$20248.1 billion) till that point; of that, Unit 1 alone had accounted for CAD1.7 billion (US$20241.2 billion) and was estimated to amount to a total cost of CAD1.98 billion (US20241.4 billion).1941

As mentioned above, three reactors (Bruce-3, Bruce-4, and Darlington-4) are currently going through this process. These projects are scheduled for completion respectively by 12 December 2026 (Bruce-3), 1 January 2028 (Bruce-4), and 23 July 2026 (Darlington-4).1942 (See Table 20.)

1. Status of Canadian Nuclear Fleet - PLEX and Expected Closures

Sources: compiled by WNISR with IESO, Operators and CNSC, 2025

Notes: OPG = Ontario Power Generation.

a – Unless otherwise mentioned, IESO, “Annual Planning Outlook - Ontario’s Electricity System Needs: 2026-2050”, April 2025.

b - IESO, “Annual Planning Outlook - A View of Ontario’s Electricity System Needs”, January 2020, see http://www.ieso.ca/-/media/Files/IESO/Document-Library/planning-forecasts/apo/Annual-Planning-Outlook-Jan2020.pdf?la=en, accessed 1 August 2020.

c - As listed on Canadian Nuclear Safety Commission’s (CNSC) website for each station, as of 24 June 2025.

Bruce: https://www.cnsc-ccsn.gc.ca/eng/reactors/power-plants/nuclear-facilities/bruce-nuclear-generating-station/index.cfm;

Darlington: https://www.cnsc-ccsn.gc.ca/eng/reactors/power-plants/nuclear-facilities/darlington-nuclear-generating-station/index.cfm;

Pickering: https://www.cnsc-ccsn.gc.ca/eng/reactors/power-plants/nuclear-facilities/pickering-nuclear-generating-station/index.cfm;

Point Lepreau: https://www.cnsc-ccsn.gc.ca/eng/reactors/power-plants/nuclear-facilities/point-lepreau-nuclear-generating-station/index.cfm.

d - Refurbishment of Bruce-6 was completed in September 2023. See Bruce Power, “Bruce Power’s Unit 6 Connected to Ontario’s Electricity Grid Following Major Component Replacement Outage”, 8 September 2023, see https://www.brucepower.com/2023/09/08/bruce-powers-unit-6-connected-to-ontarios-electricity-grid-following-major-component-replacement-outage/, accessed 26 June 2024.

e - Refurbishment of Darlington-2 was completed in June 2020, with the reactor being reconnected to the grid on 2 June 2020; see OPG, “Darlington Unit 2 Powers on—Refurbishment Now Complete on First Unit”, 4 June 2020, see https://www.opg.com/news/darlington-unit-2-powers-on/, accessed 28 July 2020.

f – IESO2025 now uses 01/09/20–05/09/23. Darlington-3 was taken offline in July 2020, with refurbishment starting in September 2020; see OPG, “Refurbishment of Darlington Nuclear’s Unit 3 Now Underway”, 3 September 2020, see https://www.opg.com/news/refurbishment-of-darlington-nuclears-unit-3-now-underway/, accessed 23 June 2025.
The refurbishment was completed in July 2023, with the reactor being reconnected to the grid on 17 July 2023.

See OPG, “OPG Celebrates the Early Completion of Darlington Unit 3”, Press Release, Ontario Power Generation, 18 July 2023, see https://www.opg.com/media_releases/opg-celebrates-the-early-completion-of-darlington-unit-3/, accessed 19 July 2023.

g - In the December 2020 issue of the IESO outlook the dates were changed to 30/07/2020–02/01/2024.

See IESO, “Annual Planning Outlook - Ontario’s Electricity System Needs: 2022-2040”, December 2020, see https://www.ieso.ca/-/media/Files/IESO/Document-Library/planning-forecasts/apo/Annual-Planning-Outlook-Dec2020.ashx, accessed 12 June 2021.

h – Provisional schedule, as presented in IESO, “Annual Planning Outlook - Ontario’s Electricity System Needs: 2026-2050”, April 2025.

i - Pickering 5–8 are planned to be taken offline in September 2026 to undergo refurbishment for an expected further lifetime extension for up to 30 years. See Government of Ontario, “Ontario Advancing Plan to Refurbish Pickering Nuclear Generating Station”, Press Release, 23 January 2025, see https://news.ontario.ca/en/release/1005620/ontario-advancing-plan-to-refurbish-pickering-nuclear-generating-station, accessed 23 June 2025.

j - OPG holds a 10-year operating license for the Pickering nuclear station, which expires on 31 August 2028. OPG is authorized to operate Units 5–8 until 31 December 2026.

k - In 2021, NB Power applied for a 25-year operating license renewal, which was granted for 10 years. Retirement is scheduled for 2044/45. See NB Power, “2023 Integrated Resource Plan: Pathways to a Net-Zero Electricity System”, 2023, see http://www.nbpower.com/en/about-us/our-energy/integrated-resource-plan/.

The third nuclear plant in Ontario at Pickering has not yet started this process. In January 2024, the Ontario Government announced that it supported refurbishing four of these reactors (Units 5–8, also referred to as “Pickering B”).1943 In October 2024, the Canadian Nuclear Safety Commission (CNSC) approved the continued operation of Pickering Nuclear’s Units 5 to 8 to the end of 2026.1944 CNSC is scheduled to hold public hearings in April and June 2026 to consider OPG’s application to refurbish Pickering Units 5 through 8.1945

Other Updates

Canada has been actively promoting Small Modular Reactors (SMRs) for around a decade (see chapter on SMRs). But over the last couple of years, a number of large nuclear reactor projects have also been proposed. These include projects at sites with existing nuclear reactors as well as new sites.

When viewed in terms of planning, the most advanced among these proposals is to build four large nuclear reactors with a capacity of 4.8 GW at the Bruce site, which already hosts eight reactors. Plans for Bruce-C can be dated back to July 2023, when Ontario’s government announced the start of pre-development work.1946 In February 2024, the federal government offered CAD50 million (US$36.5 million) for preliminary work.1947 And in June 2025, the Impact Assessment Agency of Canada and the Canadian Nuclear Safety Commission (CNSC) invited “Indigenous Nations and communities and the public to review and provide feedback on the draft Integrated Tailored Impact Statement Guidelines (draft Integrated Guidelines) and the draft Public Participation Plan (draft Plan).”1948

The anticipated time when the project would start generating power is 2045—twenty years from now.

This pace of project development is not an indication of how fast these reactors would be completed. In its initial project description, Bruce Power has laid out an “anticipated project schedule” that envisions three to four years for the impact assessment to be completed, three years after that for site preparation, and “construction and commissioning” to take “approximately 14 years”; taken together the anticipated time when the project would start generating power is 2045—twenty years from now.1949

Another proposed project with large reactors is in the Peace River area of Northern Alberta and is being proposed by a private company called Energy Alberta. As with the proposed Bruce project, the Impact Assessment Agency of Canada and the Canadian Nuclear Safety Commission invited review and feedback on the draft Integrated Tailored Impact Statement Guidelines (the draft Integrated Guidelines) and the draft Public Participation Plan (draft Plan).1950

Unlike Bruce-C, Energy Alberta envisions commissioning the first of four large reactors by 2035, with subsequent units starting operations in 2038, 2040, and 2043.1951 Energy Alberta had earlier proposed a nuclear power plant in the same region back in 2009, but then sold the proposal to Bruce Power—which eventually cancelled the project in 2011.1952

The third proposed project with large reactors is in Port Hope in Ontario; this began with a request made in January 2025 by the Ontario government to Ontario Power Generation (OPG) “to explore opportunities for new nuclear energy generation” at its Wesleyville site.1953 Some unnamed officials have been reported as saying that “the site could support” 8–10 GW of nuclear capacity, but according to the press report “the same officials declined to provide a cost estimate for such a project, saying it was too premature.”1954 Since 2018, Port Hope has been undergoing a CAD20121.28 billion (US$2012985 million) “cleanup operation” to deal with “radioactive tailings” produced while Eldorado, a former Crown corporation, refined uranium and radium between 1933 and 1988.1955 The budget was increased to CAD2.6 billion (US$1.9 billion).

Bruce Power’s initial project description listed four large reactor designs: the Monark (modified CANDU) design from AtkinsRéalis, the EPR design from Électricité de France (EDF), the Advanced Boiling Water Reactor (ABWR) design from GE-Hitachi, and Westinghouse’s AP-1000 design.1956 But the company maintains:

A reactor technology has not been selected at this time… The Impact Assessment will use a technology neutral approach, which involves the consideration of multiple technologies… Bruce Power’s evaluation of prospective nuclear technologies will focus on the value for ratepayers, opportunities for Indigenous Nations and Communities, socioeconomic benefits for the Clean Energy Frontier region of Bruce, Grey and Huron counties, as well as a number of factors including safety, reliability and cost.1957

Nevertheless, there is some reason to expect that the Monark design is a leading contender. In March 2025, the federal government “entered into a preliminary agreement with AtkinsRéalis to support the development and modernization” of the Monark design with a promise to “lend AtkinsRéalis a maximum of CAD304 million (US$220.7 million) over four years to finance half of the design project.”1958 Over the past year, former natural resources minister of New Brunswick province, Mike Holland, and the energy minister of Ontario province, Todd Smith, both joined AtkinsRéalis “within a couple of weeks” of resigning from public office, Holland as director of business development for North America in July 2024, and Smith as vice-president of marketing and business development for Atkins subsidiary Candu Energy in August 2024.1959

The Monark design was first announced in November 20231960 but the project gained momentum after AtkinsRéalis reached an agreement in February 2024 with Atomic Energy of Canada Ltd., a federal Crown corporation, “to collaborate for the purpose of successfully deploying CANDU® reactors in Canada and internationally and to expand their intellectual property licensing agreement.”1961 More recently, AtkinsRéalis chose a number of preferred suppliers, in particular, “BWX Technologies, which has manufactured major components such as steam generators for Candu plants for decades,” as summed up by The Globe and Mail.1962

Energy Alberta’s proposal for Peace River also envisions deploying up to four Monark reactors.1963

Total renewable energy capacity (incl. hydro) in Canada as of the end of 2024 amounted to 110.5 GW, up from 108.9 GW in 2023, and 96 GW in 2015.1964 The bulk of renewable capacity is hydropower which constituted 83.5 GW, up from 79.4 GW in 2015; during the same period, wind energy capacity went from 11.2 GW to 18.4 GW, and solar energy capacity more than doubled from 2.9 GW to 6.1 GW, while nuclear installed capacity (including reactors in LTO for refurbishment) dropped from 13.7 GW to 12.7 GW.

In 2024, wind turbines contributed 45.8 TWh and solar power plants contributed 5.2 TWh respectively according to data from Statistics Canada.1965 Together with hydro power, which contributed 341.8 TWh or 56 percent, renewables contributed 64.5 percent of all electrical energy supplied in Canada. As mentioned above, nuclear contributed 13.4 percent of supplied electricity.

Canadian Government targets for reducing carbon dioxide emissions, either for 2030 or 2050, place much emphasis on nuclear power.1966 The latest of the energy regulator’s Canada’s Energy Future reports that came out in 2023 did develop scenarios for a path to net zero by 2050 that projected roughly a tripling of nuclear energy generation capacity in Canada.1967 Nevertheless, these scenarios are based on unrealistic assumptions about implementation timescales and costs of potential future reactors, typically small modular reactors, that would be 2.5 to 4 times lower than estimated costs of SMRs currently in the planning.1968

Mexico

Laguna Verde nuclear power plant (LVNPP), located in Alto Lucero de Gutiérrez Barrios, in the southeastern state of Veracruz, is Mexico’s only nuclear installation. The state utility Comisión Federal de Electricidad (Federal Electricity Commission), commonly referred to as CFE, owns and operates two Mark II General Electric (GE) Boiling Water Reactors (BWRs) there.

In 2024, LVNPP generated 11.7 TWh, almost identical to the close to 11.8 TWh produced in 2023, and down from a maximum of 12.9 TWh reached in 2018.1969 It contributed 3.9 percent of the country’s electricity.

Mexico decided in the mid-1960s to build a nuclear power plant. In 1966, CFE and the National Commission for Nuclear Energy (CNEN) began the initial studies.1970 But the construction of the first reactor began only in 1976 and that of the second one in 1977.

After delays and setbacks, the first unit was connected to the grid in 1989 and the second one in 1994, each with an original design capacity of 654 MW net, a 20-year operating license, and an estimated 40-year lifespan. In 1999, the Ministry of Energy authorized an amendment to the license of Unit 2 for a 5-percent capacity increase, followed by a similar procedure for Unit 1 in 2004.

A further ~US$600 million upgrading project was launched in 2007 to increase the output of both units by an additional 20 percent. It was completed in 2011, bringing the plant’s net capacity to 1.5 GW.1971

In 2015, CFE applied for an uncommon 30-year lifetime extension to allow the reactors to operate for 60 years. In most countries, although procedures are diverse, lifetime extensions are either by 10-year periods (like in Belgium or France) or 20-year periods (like in the U.S.). In March 2019, the IAEA completed a Safety Aspects of Long-Term Operation (SALTO) review mission at the plant and made recommendations as part of the process to prepare for lifetime extension.1972 The license renewal was granted in July 2020 by the Ministry of Energy (commonly known as Sener), in the middle of the coronavirus pandemic, to allow for the operation of Unit 1 until July 2050.1973 The license for Unit 2—initially set to expire in April 2025—was extended in August 2022 by Sener, allowing the reactor to run until April 2055.1974 Reactor operation remains subject to periodic safety reviews.

Since the beginning, LVNPP has faced technical issues that have impacted its performance, like radioactive steam leaks, contaminated water discharges, and unplanned shutdowns (scrams). Therefore, the annual load factors have varied greatly, and as of year-end 2024, lifetime load factors remained below 80 percent for Unit 1 and just above 81 percent for Unit 2.1975

In addition, according to the International System on Occupational Exposure’s (ISOE) latest review, “Laguna Verde’s historical collective dose both on-line and during refueling outages, which are carried out every 18 months, is higher than the BWR average.”1976

The open dry spent-fuel storage facility (known as ISFSI) was envisaged to be fully operational in 2016 but started up two years later due to delays in the international tender organization and the delivery of the special storage containers. Then, it received fewer containers than planned according to the original scheme. It was not until 2023 that the storage facility received the 13 containers initially planned.1977

Energy Policy—Will the Sheinbaum Administration Initiate a Major Shift?

Mexico’s electricity mix is highly dependent on fossil fuels, which represented over 75 percent of the gross production in 2024 (with 62.4 percent gas, 9.5 percent oil, and 3.5 percent coal power production). Renewables accounted for 21 percent (7.2 percent solar, 6.6 percent hydro, 5.6 percent wind, and 1.7 percent geothermal and other types of production). Nuclear represented 3.5 percent of the total (gross) generation.1978

This situation is partially due to the administration of former President Andrés Manuel López Obrador (2018–2024), whose policies favored fossil fuels. His successor and protégée, President Claudia Sheinbaum, intends to provide more support for renewables and energy storage.

The Mexican government has partially reversed the 2013 energy reform that opened up the market to private investment. The counter-reform has reestablished the public companies’ hegemony, has imposed new rules for private participation, and has put the regulators, until recently independent, under governmental control. Therefore, the state is judge and player at the same time, since state-controlled regulators will oversee companies like Pemex and CFE.1979

Under the Sheinbaum administration, nuclear newbuild appears out of the picture. Even though the latest National Electric System’s Development Program 2024–2038 (PRODESEN)—like the previous 2023–2037 edition—accounts for 2.35 GW of new nuclear capacity to be deployed in 2028–2038,1980 there are no official plans to add reactors to Laguna Verde nor to build another plant (like small nuclear reactors). A few days into her presidency, in October 2024, Sheinbaum dismissed any such plans.1981 Indeed, as of mid-2025, her electricity strategy does not include more nuclear power.

In addition, CFE’s 2024–2028 business plan does not include any nuclear energy investments.1982

After Sheinbaum took office, there was a communication push by the nuclear union and columnists to develop more nuclear energy, but it failed to change her stance.1983

An issue of concern is the U.S. President Donald Trump lashing out at Mexico, more than in his previous term, focusing on issues like migration, drug policies, and bilateral trade. Even though Trump is pushing nuclear newbuild in his country, he has not stated anything regarding the deployment of the technology in Mexico. In 2018, during his first tenure in office, both countries signed a nuclear cooperation agreement1984 that entered into force in 2022 and allows for the transfer of nuclear technology and related information.1985

Sheinbaum is an environmental scientist and energy expert who used to be a researcher at the Institute of Engineering at UNAM, and a contributor to the Intergovernmental Panel on Climate Change (IPCC). She became a politician in the early 2000s, when she served as Mexico City’s Secretary of Environment under Mayor López Obrador, before acting as the metropolis’ mayor herself between 2018 and 2023.1986

Sheinbaum’s energy policy appears somewhat contradictory. On one hand, her support for Pemex and CFE means maintaining the role of fossil fuels, and her backing for natural gas infrastructure including gas-fired power plants deepens Mexico’s dependence on U.S. gas. On the other hand, her government is preparing a new climate policy—the Nationally Determined Contribution or NDC—which would take the emissions reduction target from 35 percent to 40 percent by 2030.1987 Under the previous administration, the country committed to a net-zero target by 2050, but it lacks a roadmap with middle-term goals.1988 The updated NDC is expected to be ready before COP30, to be held in the northwestern city of Belem (Brazil) in November 2025.1989

Meanwhile, pushed by gas consumption, deforestation, livestock, and agriculture, Mexico’s greenhouse gas emissions have been growing; in 2024, its emissions reached 882 million tons of CO2e according to the government.1990 The country is the second largest polluter in Latin America after Brazil and among the world’s top 15 emitters.

Laguna Verde’s Difficult Lifetime Extension

President Sheinbaum faces budgetary constraints which eventually could affect CFE’s performance and, therefore, LVNPP’s. In the first quarter of 2025, CFE registered a net loss in income of almost MXN 16.1 billion (US$836 million), reversing the positive trend from the previous three years.1991 The main drivers are debt, electricity consumption subsidies, and generation costs.1992

The future of the nuclear plant looks rather bleak, as its historical problems would keep a toll on its performance and the lack of appropriate budget appropriations might hinder proper maintenance and upgrading of the facilities. As the 2022 SALTO review pointed out, IAEA’s technical team noted that the plant still needed to:

Perform a comprehensive periodic safety review to identify potential safety improvements for LTO [Long-Term Operation].

Fully implement a programme to confirm resistance of electrical components to harsh conditions, a so-called equipment qualification programme.1993

On Mexico’s request, the IAEA did not publish the 2022 summary report, as it did in 2019, which enlists the main issues identified during the technical visit.1994

United States

See Focus Countries – United States Focus.

Asia

China

See Focus Countries – China Focus.

India

India has 21 operational nuclear power reactors, with a total net generating capacity of 7.4 GW. As of 1 July 2025, the PRIS database listed four reactors under the “suspended operation” category. However, among these, the Rajasthan-1 reactor (RAPS-1) has not generated power since 2004 and is considered permanently closed in WNISR statistics. The other three units are classified under the LTO category, with those reactors not having generated power for many years: Tarapur-1 and -2 since 2020, and Madras-1 since 2018. In December 2024, India’s Minister in charge of Atomic Energy told parliament that “TAPS-1&2 and MAPS-1 are presently under project mode”;1995 elsewhere, he has said that this “project mode” was for upgrading these reactors.1996 The Rajasthan-3 reactor (RAPS-3) that had been shut down since October 2022 for “renovation and modernisation” after 22 years of operation, was reconnected to the grid at the end of July 2024, leaving the LTO status.1997

As of mid-2025, the latest reactor to start operating is Rajasthan-7, which was connected to the grid in March 2025,1998 and declared as operating commercially in April 20251999. Construction of the reactor started with first pour of concrete in July 2011, and thus the project took 164 months to complete.2000

Reactors in India generated a record 54.7 TWh in 2024 according to IAEA-PRIS data, a surge of almost 12 percent compared to the previous year. The startup in February 2024 of the Kakrapar-4 reactor contributed alone about 30 percent of the additional production.

The share of nuclear of all the grid-fed electricity in the country, after declining since 2020, has increased slightly from 3.1 percent to 3.3 percent, returning to the same level as in 2020. However, the share calculation is at odds with the Energy Institute, which reports the same figure as PRIS for generation by nuclear reactors (54.7 TWh gross/52 TWh net), but calculates that the share of all electricity produced is a little under 2.7 percent.2001 The reason for the difference between the two sources is unclear.

Delays in Construction and Plans

India is building six more reactors with a combined net capacity of 4.8 GW. This includes the Prototype Fast Breeder Reactor (PFBR) that has been under construction since October 2004, the Pressurized Heavy Water Reactor (PHWR) Rajasthan-8, under construction since September 2011, and four VVER-1000s at Kudankulam, whose building started in June and October 2017 for the first two units, and June and December 2021 for the second pair (see previous WNISR editions).

The 500-MW PFBR is now nearly 15 years past the initially projected commissioning date of September 2010.2002 Loading of fuel into the core of the PFBR finally began on 4 March 2024.2003 In July 2024, India’s Atomic Energy Regulatory Board approved the “first approach to criticality”.2004 But there has been no announcement of the reactor reaching criticality so far, and in April 2025, the commissioning date was announced as September 2026.2005

Among the VVER reactors being imported from Russia, Kudankulam-5 and -6 are scheduled to be commissioned in 2027 and 2028 respectively, while Kudankulam-3 and -4 are now expected to be commissioned in 2026 and 2027.2006 When construction started, officials projected that these would start in March 2023 and in 2024.2007

Alongside these delays, the estimated costs of these reactors have also gone up (rounded figures hereafter). The PFBR’s cost has more than doubled from the initially anticipated Rs.35 billion to Rs.77 billion as of May 2024.2008 The total cost for Kakrapar-3 and -4 is now estimated at Rs.225 billion, up from Rs.115 billion; similarly, Rajasthan-7 and -8 were initially estimated at Rs.123 billion but the official estimate has gone up to Rs.229 billion as of December 2024. Likewise, Kudankulam-3 and -4 were initially projected to cost Rs.399 billion but the official estimate has increased to Rs.689 billion.2009

This year, again, the Indian government announced that it was planning to build a large number of PHWRs (see earlier editions of the WNISR). In March 2025, the government listed four additional reactors—GHAVP-1 & -2 in Gorakhpur, Haryana and Kaiga-5 & -6 in Karnataka—as being “projects under construction”, and eight reactors—GHAVP-3 & -4, Chutka-1 & -2, Mahi Banswara-1, -2, -3 & -4—as “projects sanctioned & under pre-project activities”.2010

Among the latter list, the four-unit Mahi Banswara Rajasthan Atomic Power Project received the siting approval from India’s Atomic Energy Regulatory Board in May 2025.2011 This follows the December 2023 government announcement that it had acquired the land for the project.2012 But construction has not started, as is the case with GHAVP-3 & -4 and Chutka-1 & -2.

More dramatic is the case of the two projects in Gorakhpur and Kaiga that have been listed as being under construction for many years now. Excavation of the site for Kaiga-5 and -6 started in April 2022.2013 For the former, in March 2018, the Nuclear Power Corporation of India Limited (NPCIL) announced that excavation of the site had commenced and that the project “is scheduled to be completed in about six years.”2014 Had that project schedule been met, the reactor would have been generating power by now. But as of mid-2025, the IAEA’s PRIS database still does not list the reactors as being “under construction”, let alone as being operational. The only significant update dates back to December 2022, when the government announced that GHAVP-1 and -2’s “foundation piles” were completed and that “testing” as well as “construction of other buildings and structures” were “underway”.2015 In June 2025, the Minister for Power reportedly announced that the first two units of the GHAVP are “expected to be commissioned by 2031” and “the remaining two by 2032”.2016

Going even further back in time, the government was explicit about when most of these reactors (GHAVP-1 & -2, Kaiga-5 & -6, Chutka-1 & -2, Mahi Banswara-1 & -2) were expected to be completed. In July 2014, the Indian government spokesperson told the parliament that GHAVP-1 & -2 were planned for completion in September 2020 and March 2021 respectively, Chutka-1 & -2 were to be completed in December 2020 and June 2021 respectively, Mahi Banswara-1 & -2 in December 2021 and June 2022 respectively, and Kaiga-5 & -6 in June 2022 and December 2022 respectively.2017 This history of lengthy delay between announcements and actual construction suggests that none of these reactors will be completed anytime soon.

The other set of projects that have been delayed even further are the reactors that were/are to be imported from the U.S. and France. These plans have been in the news ever since the negotiations of the U.S.-India nuclear deal between 2005 and 2008.2018 In 2008, Anil Kakodkar who was then the chairman of India’s Atomic Energy Commission, had envisioned importing 40 GW of light water reactors by 2020 as a result of the nuclear deal.2019

The proposal to build six AP-1000 reactors at Kovvada in Andhra Pradesh remains dormant. In February 2025, a joint statement issued by Prime Minister Modi and U.S. President Trump did not even mention this project explicitly; however, the statement did mention their “commitment to fully realize the U.S.-India 123 Civil Nuclear Agreement by moving forward with plans to work together to build U.S.-designed nuclear reactors in India through large scale localization and possible technology transfer.”2020 Two months later, when questioned in parliament about the status of the project, the Indian government could only point to acquiring the land for the reactors but explained that Westinghouse had submitted a “techno-commercial offer” in 2016. However, after Westinghouse filed for bankruptcy protection in 2017 and was subsequently taken over by Canadian holding Brookfield Business Partners, it “communicated that the earlier TCO [techno-commercial offer] submitted was no longer valid” and the government was waiting for a revised TCO.2021

In the case of the proposed nuclear reactors to be imported from France for the Jaitapur site (see past WNISR2022 and other past WNISR editions), the Indian government informed the parliament in March 2025 that “presently discussions are in progress with the French side to arrive at a viable project proposal” and “the estimated cost will emerge on conclusion of these discussions.”2022 This is more than a year after India’s Foreign Secretary announced in January 2024 that France’s EDF and NPCIL were discussing details like financing mechanisms and localization components.2023 On France’s side, as of June 2025, EDF’s website announces that a “key milestone” was reached in April 2021, when EDF’s “techno-commercial binding offer [was] submitted” and that “NPCIL and EDF [would] work alongside in order to reach a General Framework Agreement in the coming months.”2024 That did not happen since.

Despite these lengthy delays, the government continues to set ambitious goals—a practice that dates back to the 1950s.2025 In the 2025–2026 budget presentation, the government announced a new one: “at least 100 GW of nuclear energy by 2047” including “at least 5 indigenously developed SMRs [Small Modular Reactors]…operationalized by 2033.”2026 The significance of 2047 is that it would mark 100 years since India’s independence from British colonialism. An earlier target from 1999 also had a similar rhythm to it: 20,000 MW by 2020.2027 As detailed in WNISR2020, India’s net capacity was just 6,200 MW, less than a third of the target for that year.

When the government announced the latest target, the Chairman of the Atomic Energy Commission told the media that 100 GW by 2047 is “very achievable” but “getting land and appropriate nuclear fuel could be a limiting factor” while the “private sector can help augment capacity by building nuclear plants especially small modular reactors.”2028 The mention of land being a limiting factor testifies to the intensity of opposition movements that have sprung up near all nuclear reactor sites.2029

Whether involving the private sector is going to change anything significantly about the future of nuclear power in India remains to be seen. However, the government is reportedly considering amending laws governing the nuclear sector in India to ease participation of private sector companies.2030 This is not entirely new; for example, in November 2022, the Minister responsible for Atomic Energy invited private sector companies to participate in building SMRs.2031 In July 2024, Tata Power, part of a large industrial conglomerate in India, announced that it will “explore participation in Small Modular Nuclear Reactors, once the Government gives necessary permissions…”2032 In the meanwhile, the company announced in its latest annual report that it is “preparing proactively by evaluating potential sites, securing water resources, and reviewing advanced nuclear technologies, including small modular reactors.”2033

Another part of the same conglomerate, Tata Consulting Engineers signed an agreement with GE-Hitachi in 2010 to “explore potential project design and workforce development opportunities in support of GEH’s future nuclear projects in India and around the world.”2034 But nothing seems to have resulted from that agreement. According to media reports, the Department of Atomic Energy expects Public-Private Partnerships to account for around 50 percent of the 100 GW target.2035

Power Mix

Renewable energy continues to progress rapidly in India. According to the International Renewable Energy Agency (IRENA), total renewable energy capacity (including large hydro plants) grew by a factor of 2.6 from 78.6 GW in 2015 to 204.3 GW in 2024.2036

In 2024, solar energy totals 97.4 GW with a 33.7 percent increase over 2023. Wind energy capacity reached 48.2 GW, an increase of 7.7 percent compared to 2023. With the grid connection of one new reactor, nuclear capacity increased by 0.6 GW. According to the Central Electricity Authority, the official data source on India’s electricity sector, the capacity of non-hydro modern renewables was 178.8 GW with hydropower at 47.9 GW as of 31 May 2025.2037 Coal continues to dominate with an installed capacity of 212.7 GW, also as of 31 May 2025, whereas the installed capacity of gas is 20.1 GW. But, in July 2025, the government announced that 50 percent of India’s installed capacity was now non-fossil fueled, 242.8 GW of 484.8 GW.2038 In 2015, as part of its commitments to the Paris Agreement, the Indian government declared a target of non-fossil fueled sources contributing 40 percent of the total generation capacity by 2030.2039

For 2024, the Energy Institute reports the combined output of modern renewables (excluding large hydropower) in India as 240.5 TWh gross—up from 221.2 TWh in 2023 (revised from the 232.8 TWh figure listed by the Energy Institute last year)—representing 11.8 percent of all the electrical energy.2040 Of this, wind energy contributed 81.5 TWh and solar energy 136.8 TWh. Since nuclear energy contributed 54.7 TWh, wind and solar power together produced nearly four times the amount of electricity that nuclear reactors produced. However, fossil fuels together generated 1,577 TWh, nearly 78 percent of all electricity generated in 2024, with coal alone comprising three-quarters of the total.

Japan

See Focus Countries – Japan Focus.

Pakistan

Pakistan operates six nuclear reactors with a combined (net) capacity of 3.3 GW. Nuclear electricity production has increased from 21.3 TWh in 2023 to a new all-time high of 21.7 TWh net in 2024, while the share of electricity from nuclear power plants increased from the 16.2-percent peak in 2023 to a record 17 percent in 2024.

All operating reactors were built by the China National Nuclear Corporation (CNNC). This includes two Hualong One reactors (Kanupp-2 and Kanupp-3) outside the city of Karachi and four CNP-300 nuclear reactors in Chashma. CNNC is also building another 1200-MW Hualong One reactor in Chashma (Unit 5). The agreement to build this reactor dates back to 2017,2041 but it took over seven years to progress to the formal construction start, i.e., first pour of concrete for the base slab of the reactor building, which occurred on 30 December 20242042. At the function, Pakistan’s Minister for Planning, Development, and Special Initiatives said that construction “reflected the strength of Pakistan-China cooperation.”2043 It is also China’s only ongoing nuclear newbuild project abroad and represents the first non-Russian construction start anywhere in the world in the past five years.

In January 2025, the National Electric Power Regulatory Authority (NEPRA) published an estimated overnight cost of PKR966 billion (US$3.4 billion) for the Chasma-5 (CHASNUPP-5 or C-5) project and total cost (including financing and other costs) at PKR1,125 billion (US$4 billion).2044 The majority of the cost is planned to be covered by credit from China.2045 As construction started, the reactor was reportedly projected to start producing power by 2030, which echoed earlier official announcements.2046 The project has been criticized for its high cost of power, and shelving renewable energy projects to make way for it.2047

Pakistan’s renewable electricity capacity was 15.2 GW in 2024, up from 14.2 GW in 2023.2048 While hydropower, with a total capacity of 11.5 GW, was the most important component of this capacity, solar energy is the fastest growing source of energy. In 2024, the total capacity of solar energy was 1.4 GW, up from 1.2 GW at the end of 2023, while wind constituted 1.8 GW, same as the two previous years. The reported solar energy capacity likely does not reflect the full deployment of the technology in Pakistan, because there are reports of importing 22 GW of solar panels from foreign suppliers.2049

Renewable energy sources (including hydro) produced 44.5 TWh net in 2024, comparable to the 2023 figure of 44.4 TWh; wind and solar energy contributed 3.4 TWh and 1.6 TWh respectively in 2024.2050

South Korea

See Focus Countries – South Korea Focus.

Taiwan

See Focus Countries – Taiwan Focus.

Middle East

Iran

Iran only operational reactor, Bushehr-1—also spelled Busheer-1—has a net capacity of 915 MW and was connected to the grid in September 2011.2051 IAEA-PRIS statistics for this reactor shows a record production of 6.4 TWh in 2024. Per the Energy Institute, with 7.3 TWh gross, nuclear covered a stable 1.9 percent of the total electricity generated in the country.

A second unit with a net capacity of 974 MW, also a Pressurized Water Reactor from Rosatom, had been on the IAEA’s list as under construction for almost 20 years before vanishing in 2005. It has been relisted as under construction since September 2019 in the IAEA’s PRIS database. However, the first pour of concrete for the foundations of the reactor building, the traditional marker for official start of construction, only happened in November 2019 according to the Atomic Energy Organization of Iran (AEOI).2052

The contract for the second unit was signed as part of a two-unit project, Bushehr-2 and -3, by the Nuclear Power Production and Development Company of Iran and Rosatom’s subsidiary Atomstroyexport JSC back in November 2014.2053 These two units were to be added “in the short-term”; Rosatom also announced that two more reactors would be built at Bushehr “in the medium-term”, and that it would develop “an entirely new nuclear power plant in the country with four VVER reactors” at a location that was “yet to be announced by the Iranian nuclear authorities.”2054

Apart from Bushehr-2, no official construction start has been reported, and WNISR does not consider any other project as under construction. However, Iranian officials routinely talk about “Units 2 and 3” being under construction. In February 2025, for example, the AEOI website featured a press release that announced that Iran’s President Dr. M. Pezeshkian, “inspected the progress of the construction of units 2 and 3 of the Bushehr nuclear power plant.”2055

At the end of 2019, these two reactors were expected to be completed in 2024 and 2026, respectively.2056 However, in September 2024, at a side event of the IAEA’s General Conference, officials were reported as saying that the reactor pressure vessel for Unit 2 was to be installed “30 months later” (namely, in May 2027), and that the “physical start-up” was anticipated to be in 2029.2057

Some work is under way, according to the AEOI, for a domestically designed 300-MW reactor, at the Darkhovin or Karun nuclear power plant, in Ahvaz, Khuzestan province2058 and for four Russian-designed reactors of 1250 MW each, in Sirik, Hormozgan province.2059 Both of these projects were officially announced in 2024. AEOI has also identified Shiraz as another potential location for a nuclear power plant.2060 The official goal is to expand nuclear capacity to 20 GW by 2040.2061

On 9 June 2025, AEOI’s President Mohammad Eslami reportedly confirmed that an agreement had been signed with Russia on the construction of eight reactors, it being understood that this includes four units at the Busheer site—only one of which is under construction—and the four units at Darkhovin. Russia is supposed to provide the funding for the projects.2062

Iran is highly reliant on fossil fuels for electricity generation—mainly gas, which accounted for 86 percent of the total gross production—with renewable energy, including hydro, accounting for only 5.3 percent, down from 6.5 percent in 2023.2063 There are plans to increase wind energy capacity by 5.5 times in line with the 7th Development Plan’s target of 30 GW by 2028.2064 Furthermore, Iran is planning to add 15 GW of solar capacity.2065 While no explicit deadline has been set to achieve this target, in April 2023, the Iranian Ministry of Energy announced that 10 GW of renewable energy capacity would be added by August 2025.2066

United Arab Emirates

The United Arab Emirates (UAE) operates four South Korean APR-1400 reactors at the Barakah nuclear project with a total (net) capacity of 5.3 GW. The last of these reactors was connected to the grid on 23 March 2024,2067 and entered commercial operation in September 20242068. The earlier three units were declared entering commercial operation in April 2021, March 2022, and February 2023 respectively.2069 Each of these was about four years delayed compared to the timeline projected by the Emirates Nuclear Energy Corporation (ENEC) in 2014.2070

As reported in WNISR2024, ENEC is considering building more nuclear reactors, and launched the so-called ADVANCE program in late 2023 to “evaluate Small Modular Reactors (SMRs) and so-called Advanced Reactor technologies for both domestic and international investment and development.”2071 In May 2025, it signed an MoU with General Electric Vernova Hitachi to “evaluate the deployment of the BWRX-300 Small Modular Reactor (SMR) technology internationally.”2072 This follows an earlier agreement signed with GE-Hitachi, as well as five other SMR vendors, on the sidelines of the 28th Conference of the Parties of the UN Framework Convention on Climate Change (COP28) in 2023.2073 In addition, in early 2025, ENEC signed an agreement with Newcleo “to explore opportunities to work together to support the advancement of LFR [lead cooled fast reactor] projects.”2074 All of these are non-binding agreements.

In 2024, nuclear reactors contributed 38.6 TWh (net) of electricity to the UAE’s grid, or 22.9 percent of the total electricity generation (gross). The corresponding figures for 2023 are 32.7 TWh and 20.6 percent. Natural gas remains the main source of electricity generation in the country, contributing over 68 percent of the total.2075 Renewables contributed 15.5 TWh net, or 8.8 percent of all electricity supplied to UAE’s grid.2076

When compared to earlier years, installed capacity of renewables grew relatively slowly, by less than 1 percent, in 2024 to a total of 6.14 GW; in comparison, the capacity in 2022 was only 3.6 GW.2077 Solar energy dominates with 6 GW, including photovoltaics (5.4 GW) and concentrated solar power (600 MW). According to the 2023 update of its energy strategy, by 2030, the UAE plans to triple the capacity of renewable energy compared to 2022 and generate 32 percent of its power by nuclear and renewables.2078

European Union (EU27)

The EU27 member states have gone through three nuclear construction waves (see Figure 80)—two small ones in the 1960s and the 1970s and a larger one in the 1980s (mainly in France). Over the past 30 years, 1995–2024, only 13 reactors were connected to the grid in current EU27 Member States, six of them in Western Europe—five in France and one in Finland—and seven in Eastern and Central Europe—three in Slovakia, and two each in the Czech Republic and Romania. Only four reactors started up in the past 20 years, 2005–2024: after Cernavoda-2 was connected to the grid in Romania in 2007, it took another 15 years, until the following reactor, Olkiluto-3 in Finland, produced its first kilowatt-hours in March 2022; Mochovce-3 in Slovakia, where construction first started in 1985, was finally connected to the grid in January 2023; and Flamanville-3 in France started up in December 2024.

As Figure 81 shows, 100 reactors are operating in the EU27 as of mid-2025, 36 less than the historic maximum in 1989, a drop by over one quarter. Eighty percent of the operating plants are located in six of the western countries—with 57 units in France alone—and only 20 in six newer member states.

1. Nuclear Reactors Startups and Closures in the EU27, 1959–1 July 2025

Sources: WNISR with IAEA-PRIS, 2025

The closures of Tihange-2 and Doel-1 in Belgium in February 2023 and February 2025 respectively, and Emsland, Isar-2, and Neckarwestheim-2 in Germany in April 2023, brought the number of permanently closed reactors in the EU27 to 78 (69 in Western Europe, over half of which in Germany). Thirty-six units were closed over the past 20 years from 2005 to mid-2025.

1. Nuclear Reactors and Net Operating Capacity in the EU27

Sources: WNISR with IAEA-PRIS, 2025

In 2024, nuclear plants generated 617 TWh net in the EU27, an increase of almost 5 percent compared to the previous year; production had increased by almost 2 percent in 2023, following a plunge of 17 percent in 2022. Consequently, the nuclear share in gross power generation had dropped to 21.7 percent in 2022, slightly increasing to 22.6 percent and 23.2 percent in 2023 and 2024 respectively.

In the absence of any significant delivering newbuild program (see Figure 82), the average age of nuclear power plants has increased since the mid-80s and at mid-2025 is 38.7 years (see Figure 83 and Figure 84). Following grid connection of Olkiluoto-3, Mochovce-3, and Flamanville-3, only one reactor remains under construction in the EU27, Mochovce-4, in Slovakia, where construction originally began 40 years ago, in 1985.

When the last unit has finally been connected to the grid, in late 2025 or early 2026 (?), in view of the long lead times of nuclear construction projects, there will be no new nuclear power reactors starting up in the EU27 for years to come. In other words, the current nuclear fleet will continue aging and shrinking at least in the medium term.

1. Construction Starts of Nuclear Reactors in the EU27

Sources: WNISR with IAEA-PRIS, 2025

Note: Construction start of Mochovce-3 and -4 was first introduced as of 1985 in IAEA-PRIS, “Nuclear Power Reactors in the World – April 1986 Edition”, 1986. Their construction was later suspended. See section on Slovakia.

1. Age Evolution of EU27 Reactor Fleet, 1959–2024

Sources: WNISR with IAEA-PRIS, 2025

The age distribution shows that now almost nine in ten (87) of the E.U.’s operating nuclear reactors have been in operation for 31 years and beyond of which 42 have been on the grid for 41 years and more. One reactor, Borssele in the Netherlands, the oldest in the E.U., reached 51 years of operation in July 2024. The then second oldest reactor, Doel-1, was closed after just over 50 years of operation in early 2025.

1. Age Distribution of the EU27 Reactor Fleet

Sources: WNISR with IAEA-PRIS, 2025

In June 2025, the European Commission has published the long-awaited Nuclear Illustrative Program (PINC). It estimates that “Delivering Member States’ plans regarding nuclear energy will require significant investments, of around €241 billion (US$274 billion) until 2050, both for lifetime extensions of existing reactors and the construction of new large-scale reactors,” not including SMRs or other types of “advanced reactors”. The report projects installed nuclear capacity across the E.U. is to grow “from 98 GWe in 2025 to 109 around GWe by 2050”.2079 In other words, to hardly grow over the 25-year period. However, previously, the Commission envisaged nuclear capacity to continue the declining trend.

The European Commission does not carry out any feasibility assessment: “It aggregates Member States’ plans for nuclear energy as declared in their National Energy and Climate Plans (NECPs) and derives the evolution of installed generation capacity from large-scale reactors, as well as the associated investment needs.” The Commission also states in a “Working Document” distributed alongside the communication on PINC—wrongly—that besides the Mohovce-4 unit, two reactors would be under construction at the Paks site in Hungary although none of the units have achieved the milestone of concreting the base slab of the reactor building, the official construction start.

The Commission nevertheless carried out a sensitivity analysis of potential delays judging an upper limit of 10 years delay credible. Such delays “could shift the installed capacity in 2050 by more that 20 GWe”, which could drop to 85 GWe by 2050. The Commission recognizes that

The EU’s nuclear industry is facing significant challenges in responding to current and future nuclear development trends across the segments including new builds and lifetime extensions, (as well as decommissioning and waste management). Therefore, the industry has to undergo a substantial transformation to increase its capacity, effectiveness, and competitiveness.2080

and surprisingly bluntly describes the nuclear industry’s attractiveness issue:

[The] Nuclear sector has an image problem and is not attractive to young people, often not even considered as an option.

Western Europe

As of mid-2025, 93 nuclear power reactors operated in Western Europe (western E.U. member states, U.K., and Switzerland), 67 fewer units than in the peak years 1988–89, a 42-percent decline. Over the year, Flamanville-3 started up and Doel-1 closed.

With Switzerland operating two reactors for over 50 years—Beznau-1 at 56, and Beznau-2 at almost 54)—the average age of operating reactors in Western Europe reaches 40.3 years (see Figure 85).

1. Age Distribution of the Western European Reactor Fleet (incl. Switzerland and the U.K.)

Sources: WNISR with IAEA-PRIS, 2025

The only remaining active construction project is implemented by French utility EDF at Hinkley Point in the U.K., where two European Pressurized Water Reactors (EPR), Hinkley Point C-1 and C-2, have been under construction since 2018 and 2019 respectively. As previous EPR projects in Finland and France, both units are billions over budget, years behind schedule, and are currently not expected to be commissioned before around 2030 or later (see United Kingdom Focus).

The mean-age evolution of the nuclear reactor fleet in Western Europe follows the same pattern as the EU27, constantly increasing since the middle of the 1980s. The eventual startup of the two reactors currently under construction will have little impact on the aging aspect of an otherwise shrinking fleet.

Belgium

From an original fleet of seven reactors, as of 1 July 2025, four reactors (two at Doel and two at Tihange) totaling 3.5 GW remained connected to the Belgian grid. The 445-MW PWR Doel-1 was closed on 15 February 2025 in line with a nuclear power plant phaseout plan that had come into effect in 2003 (see Table 21). This legislation limited the operational lifetime of Belgian commercial power plants to 40 years and prohibited the construction of additional reactors.2081 In 2021, the government of Alexander De Croo confirmed the final closure year for all Belgian reactors as 2025.2082 But in 2022, following the Russian invasion of Ukraine and subsequent energy supply challenges, the government initiated preparatory steps to extend the lifetimes of the youngest, 40-year-old units Doel-4 and Tihange-3 beyond the 2025 closure date.2083 The current average age of the four units is 45 years.

Negotiations are ongoing between operator Electrabel, owned by French Engie, and the Belgian government on the details of the conditions for lifetime extensions. In mid-2023, the parties signed an interim agreement to extend the lifetimes of Doel-4 and Tihange-3 by ten years. For this, the Belgian state would become equal co-owner of the plants with Engie and implement a “contract for difference model with incentives for the industrial operator to achieve [favorable] technical and economic performance at the plants.” The state would also receive a €15 billion (US$202215.8billion) lump sum payment from Engie, ridding the company of all “future costs related to the treatment of nuclear waste” not only for the two lifetime-extended units but for all seven Belgian power reactors, both operational and closed.2084 Additionally, according to the deal, the Belgian state and Electrabel will provide 89.807 percent of the €2–2.5 billion (US$2.3–2.8 billion) total capital contribution to fund the cost of the lifetime extension backfitting and upgrading work; the remainder will be covered by Electricité de France (EDF) subsidiary Luminus that owns shares of the Doel plant. As of February 2025, the exact composition of the funding, i.e., the distribution of shareholder loans and equity injection, was yet to be decided.2085 The new law allowing for the lifetime extension of Doel-4 and Tihange-3 was enacted on 26 April 2024 and promulgated on 5 June 2024.2086 See Belgium Focus in WNISR2024 for further details on the procedure.

The European Commission approved the deal on 21 February 2025 after having carried out an investigation, launched in July 2024, “to assess whether public support that Belgium plans to grant for the lifetime extension of two nuclear reactors (Doel 4 and Tihange 3) is in line with EU State aid rules.”2087

The government and Engie officially signed the deal on 14 March 2025, paving the way for the remaining preparatory steps to operate Doel-4 and Tihange-3 until 2035 (see Figure 86). For the operator, these include organizing fuel deliveries, receiving a final approval (expected in June 2025) from the Belgian nuclear regulatory agency FANC/AFCN,2088 and completing engineering tasks during the scheduled temporary shutdown of the reactors—from April to July and from July to October 2025 at Tihange-3 and Doel-4, respectively.2089 The first installment of €202211.5 billion (US$202212 billion) was subsequently paid to the specially set-up agency Hedera. The remainder of the lump sum payment is expected to be transferred once both reactors have returned to operations in November 2025.2090

The Belgian electricity mix is still dominated by nuclear power. In 2024, the nuclear fleet produced 31.4 TWh gross, down by 1.5 TWh compared to 2023, and corresponding to 41.9 percent of Belgian electricity. The remainder was provided by natural gas (18.7 percent), solar (11.9 percent), wind (18.7 percent—offshore 9.6 percent and onshore 9.1 percent), bioenergy (4.7 percent), and hydro (0.6 percent).2091 The historic maximum nuclear share was 67.1 percent in 1986.

1. Belgian Nuclear Fleet (as of 1 July 2025)

Reactor	Net Capacity (MW)	Grid Connection	Operating Age (as of 1 July 2025) / Age at Closure	End of License (Planned or Actual Closure Date)
Doel-1	433	28/08/1974	50.5	15 February 2025 (Closed on 14 February 2025)
Doel-2	433	21/08/1975	49.9	10-year lifetime extension to 1 December 2025
Doel-3	1 006	23/06/1982	40.3	1 October 2022 (Closed on 23 September 2022)
Doel-4	1 038	08/04/1985	40.2	10-year lifetime extension (Closure date 2035–2037)
Tihange-1	962	07/03/1975	50.3	10-year lifetime extension to 1 October 2025
Tihange-2	1 008	13/10/1982	40.3	1 February 2023 (Closed on 31 January 2023
Tihange-3	1 038	15/06/1985	40.0	10-year lifetime extension (Closure date 2035–2037)

Sources: Various, compiled by WNISR with Belgian Laws of 28 June 20152092 and 26 April 20242093

With the closure of Doel-1 in February 2025 and the expected closures of Tihange-1 and Doel-2 in October and December 2025, respectively,2094 nuclear power production is bound to decrease. Most recent Belgian energy policy envisioned carbon neutrality by 2050, to be achieved mostly through ambitious renewables expansion targets, including large offshore wind projects. Belgium missed the June 2024 deadline to submit its final updated National Energy and Climate Plan (NECP) to the European Commission, which in turn launched an infringement procedure against the country in October 2024.2095

Federal elections held in June 2024 resulted in a new government, dubbed the “Arizona coalition”, that comprises five partners and is led by Flemish nationalist Bart De Wever, who was sworn into office on 4 February 2025.2096 As of 30 June 2025, the government had still not provided a final updated National Energy and Climate Plan (NECP), initially due in June 2024.2097

1. 80-Year Histogram of Belgian Nuclear Program

Sources: WNISR with IAEA-PRIS and Belgian Law of 26 April 2024, 2025

Note: According to the Belgian law of 26 April 2024, Doel-4 and Tihange-3 can operate for an additional period of 10 years from the restart date, on the understanding that nuclear power plants will be deactivated at the end of this period and no later than 31 December 2037.

The resulting February 2025 coalition agreement promises “enormous transformations” of Belgium’s energy landscape involving numerous ambitions for the nuclear sector, including by securing 4 GW of “existing” nuclear in the short term, and “striv[ing] to remove all obstacles, and to facilitate and accelerate the construction of new nuclear reactors.” The agreement includes the following:

• “in the very short term, the government will take all necessary measures to extend the operating life of existing units that meet safety standards and will begin discussions with the nuclear operator and owners to this effect”—it is somewhat unclear which units this refers to, the ones still operating, the ones already slated for lifetime extension, or the ones already closed
• “the government will launch an ambitious program in the short term to relaunch the nuclear industry in Belgium and build new nuclear reactors”
• the government will lift all legal dispositions relative to the nuclear phaseout and the prohibition of the construction of new reactors
• concerning Doel-4 and Tihange-3, “the government opts for a minimum additional lifetime extension of 10 years”
• concerning SMRs, the government recommends “the introduction of a standard certificate at European level and the shortening of licensing procedures.”2098

In March 2025, the newly appointed Energy Minister Mathieu Bihet explained in an interview with the newspaper De Tijd that the new coalition’s energy policy was “based on two pillars: nuclear energy and renewable energy.”2099 In an energy policy brief, Bihet announced that “taking into account […] sustainability, safety, and cost optimization, the government [aimed] to integrate a capacity of 4 GW of nuclear energy in the electricity mix” without providing a target year or any further detail.2100

The Belgian nuclear operator has a different view of the potential options. In January 2025, the CEO of Engie Belgium, Vincent Verbeke, told Belgian news agency VRT NWS that further extending the lifetimes of Doel-4 and Tihange-3 was “unthinkable” and that nuclear was no longer part of Engie’s “strategic ambition.”2101

The political shift towards nuclear power culminated in a parliamentary vote after which the phaseout legislation was scrapped and the ban on constructing new nuclear power plants was lifted.2102 On 15 May 2025, the legislation passed with a large majority (102 in favor, eight against (Ecolo-Groen), and 31 abstentions (Socialist Party and far-left Worker’s Party PTB).2103 The current Arizona coalition has only 80 seats.2104

Now that legislative hurdles have been removed, the government has to concretize its plans and convince the Belgian nuclear operator for nuclear power’s relaunch in the country. How this pans out, and whether new nuclear power plants will be seriously pursued, remains to be seen. As of May 2025, there were conflicting statements and reports regarding Engie’s enthusiasm towards new nuclear or further lifetime extensions.2105 Luminus, majority owned by EDF (68.6 percent), is a shareholder in the Belgian reactors Doel-4 and Tihange-3 (10.2 percent of each), while EDF Belgium, a full EDF subsidiary, co-owns Tihange-1 with Electrabel (50 percent each).2106 In March 2025, Luminus CEO Grégoire Dallemagne expressed the company’s interest in cooperating with the Belgian government on the expansion plans, with the company also planning to invest up to €4 billion (US$4.71 billion) into energy projects, including renewables and batteries, by 2030.2107

Meanwhile, Belgian renewables capacities have substantially increased over the past decade. From 2015 to 2024, onshore wind capacity grew from 1.5 GW to 3.3 GW, offshore wind more than tripled from 712 MW to an estimated 2.26 GW, and solar PV grew by a factor of 3.1 to over 9.7 GW of installed capacity. However, year-on-year onshore wind power capacity expansion was only 145 MW, from approximately 3.2 GW in 2023 to just above 3.3 GW in 2024, while offshore capacities have remained stagnant at 2.26 GW since 2020.2108 Plans for substantial offshore wind capacity expansion in the “Princess Elisabeth” zone, about 45 km off the Belgian coast, are advancing with the first 700-MW tender launched at the end of 2024. As of November 2024, Belgium planned a total of 5.8 GW of offshore capacity on its grid by 2030.2109

Finland

Nuclear reactors supplied 31.1 TWh of electricity in Finland in 2024, down 5.2 percent from 2023. This represented a 39.2 percent share in the electricity mix, a significant drop from the peak of 42.1 percent in 2023.

Finland has adopted different nuclear technologies from various suppliers. Two of its operating reactors are modified VVER-V-213 units (507 MW each) built by Russian contractors at Loviisa and two others are AAIII BWR-2500 models (890 MW each) built by Asea Brown Boveri (ABB) at Olkiluoto. These four units were built in the 1970s.

With the deployment of the 1575-MW European Pressurized Water Reactor (EPR) Olkiluoto-3 (OL3), a third generation (Gen III) technology was added to the fleet. The project was completed with a 13-year delay by French-German consortium Areva-Siemens in 2022 and provided electricity at full capacity for the first time on 16 April 2023, 17.5 years after construction began.2110 The full operational transfer, and with it the cessation of joint liabilities between operator Teollisuuden Voima Oyj (TVO) and Areva-Siemens, was completed in June 2025.2111 Refer to previous WNISR editions for detailed accounts of the OL3 construction project.

Fluctuating Nuclear Production

In 2024, with 96 percent of electricity demand covered by domestic production,2112 Finnish electricity generation fell just short of the government-envisioned target of electrical self-sufficiency, partially due to several unplanned and extended outages at its nuclear power plants.2113 The outages were not quite compensated for by a record year of wind production with over 20 TWh.2114 The high share of domestic electricity production helped stabilize the Finnish grid2115 which had to operate with reduced interconnection capacity since the EstLink 2 line that links Estonia to Finland was damaged on 25 December 2024 possibly by a tanker anchor, after having returned to service only in September 2024 following an outage of over seven months.2116 The transmission line was returned to operations in June 2025.2117

Lower-than-planned output from OL3 was caused in part by delays in the first annual maintenance outage from which the plant returned five weeks late in May 2024, taking twice as long as planned “as a result of technical problems identified during maintenance activities” per the operator TVO.2118 Unexpected outages and capacity reductions continued throughout the year. For example, in October 2024, the reactor operated at a reduced power level after a control rod was “unexpectedly dropped into the reactor,”2119 and in November 2024, a turbine malfunction interrupted electricity generation,2120 which caused a massive frequency loss in the Nordic power grid.2121 In a LinkedIn post listing OL3’s malfunctions that caused substantial frequency changes in the grid in 2024, the Danish battery service provider Hybrid Greentech provocatively stated that “Olkiluoto is starting to compete with the Swedish nuclear power plant Forsmark for being the leading cause of major (loss of generation) disturbances in the Nordic power system” (see Sweden in Annex 1).2122

Furthermore, since the start of its operations, OL3, as a participant in the Nordic electricity regulation market, had to reduce its output on several occasions due to high wind generation2123 and grid constraints (lack of reserve capacity).2124 In 2025, the annual maintenance outage was completed in just under 60 days instead of the planned 63 days, with operator TVO hoping that past malfunctions and unreliability had resulted from an initial lack of experience.2125 Nevertheless, during the outage, the improper closure of a hatch—attributed to human error—resulted in about 100 cubic meter of coolant flowing into “containment rooms closed to the environment and into the floor drain system of the containment.” TVO stated that “the incident did not pose any risk to the personnel, the environment, or nuclear safety.”2126

OL2 was taken off the grid in early September 2024, when concerning moisture levels were detected in the plant’s generator. A fault in one of the rotors of this water-cooled generator was identified as the cause.2127 The faulty equipment was replaced with a spare rotor. After 28 days, OL2 was reconnected to the grid, albeit at reduced capacity—725 MW instead of 890 MW (net)—to “prevent further damage and due to a shortage of spare parts,” according to a subsequent statement from TVO.2128 Nevertheless, in April 2025, while in the process of bringing the unit back to full power—apparently made possible by “the arrival of a replacement rotor at Olkiluoto”—the same issue reemerged, and the reactor was shut down again on 15 April 2025.2129 The replacement rotor was subsequently installed. It was a refurbished spare part, of which there were none left in TVO’s inventory. The unit was eventually reconnected to the grid in early May 2025, a few days ahead of the initial schedule.2130 OL2 will operate at reduced capacity (735 MW) until its next annual outage in May–June 2026.2131

At OL1, operations continued as scheduled until 21 May 2025, when a malfunction at the turbine island disrupted ramp-up operations following the unit’s annual maintenance outage. The reactor returned to the grid on 24 May 2025.2132

At the Loviisa plant, both reactors underwent annual planned maintenance. At Unit 1, an extensive outage, which takes place every four years, was extended by nine days for additional maintenance work, and at Unit 2, the shorter yearly outage concluded two days later than planned.2133 Technical issues with the cooling system of the reactor’s control rod drive mechanisms at Loviisa-2 led to an unexpected shutdown on 18 November 2024, the same week OL3 experienced an unplanned production outage (see above).2134 Loviisa-2 was reconnected to the grid four days later.2135 The same unit’s output was halved for a few days in May 2025 when a hydrogen leak was detected at one of the steam generators.2136

Finnish Reactor Lifetime Extension Plans

As of mid-2025, the average age of the four reactors subject to lifetime extensions (OL1, OL2, Loviisa-1, and -2) was 46.3 years. In 2018, OL1 and OL2 (connected to the grid in 1978 and 1980, respectively) were approved to operate until 2038, allowing for up to 60 years of operation.2137 In December 2024, operator TVO submitted an Environmental Impact Assessment (EIA) report evaluating the potential consequences of further power uprating at OL1 and OL2—by about eight percent, to 970 MW each—and extending the reactors’ operating licenses to 2048 or even 2058. According to TVO documentation, an application for the uprating and potential lifetime extensions would have to be submitted to the Finnish regulator STUK before the end of 2028 if the EIA report is deemed appropriate by the Ministry of Economic Affairs and Employment.2138

In April 2025, the Ministry concluded that the report conformed with current legislation,2139 prompting TVO to take on a €75 million (US$81 million) loan from the Nordic Investment Bank. The “considerable” investment project is presented as primarily “aimed at maintaining and improving safety measures”, through the replacement at both units of the Instrumentation & Control (I&C) and monitoring systems and modernization of the steam separators. However, possible lifetime extensions and uprates—for which TVO says it is preparing a “decision-making process”—have explicitly “been accounted for in the investment plans.”2140

In February 2023, both 507-MW Russian-built reactors at the Loviisa plant were granted operational lifetime extensions until the end of 2050, resulting in 73 and 70 years of planned operation for Loviisa-1 and -2, respectively.2141 Beginning in 2026, work on the modernization of the low-pressure turbines shall increase the total plant capacity by around 38 MW.2142 In February 2023, the company indicated that after having already invested €300 million (US$2023324 million) into refurbishment over the previous five years,2143 it estimates that by 2050 a total of €1 billion (US$20231.1 billion) will have to be invested to continue operation. The company reiterated the estimate in 2024 and early 2025.2144

Fuel Supply Diversification

The Loviisa operator, Fortum, is diversifying its fuel supply away from Russian supplier TVEL and loaded the first batch of Westinghouse-manufactured fuel into Loviisa-2 in August 2024.2145 More fuel is expected to be delivered in 2025 by Spanish fuel rod manufacturer Enusa, which signed a ten-year cooperation deal with Westinghouse earlier this year.2146

New Nuclear Projects and Current Power Mix

In May 2025, Russian state-owned vendor Rosatom filed a US$2.8 billion lawsuit in Moscow against Fortum and steel manufacturer Outokumpu, who, together with SSAB and other Finnish companies, had held two thirds of the Fennovoima project company (one-third held by Rosatom) that was set up to manage the Hanhikivi power plant project in 2013. The project was canceled unilaterally by the Finnish contractors in the wake of the Russian invasion of Ukraine in 2022. Refer to previous WNISR editions for details on the Hanhikivi project and its cancellation.

Since then, several litigation claims have been made by all parties that are mostly under confidential arbitration at international courts. Outokompu dismissed Rosatom’s most recent reported filing by stating that Moscow was “not the appropriate venue to address the related disputes.”2147 In another instance, the Finnish customs agency commenced an investigation against an “operator of a nuclear power plant construction project” who allegedly attempted to export archival material of the project to Russia, which is currently forbidden under E.U. sanctions.2148

Despite the challenges of the OL3 construction project and the cancellation of the Hanhikivi plant, some Finnish utilities are envisioning further nuclear power expansion. This is supported by the government, with Juho Korteniemi, Head of the Energy Department’s Nuclear Energy Unit at the Ministry of Economics, stating that the “road is […] open for new projects […].” However, perhaps preemptively excluding Russian vendors, he added that the “applicant’s background [must be] acceptable from the point of view of national security.”2149

In January 2025, then-Environment Minister Kai Mykkänen2150 said that “the Finnish state should invest in a new, large nuclear power plant” because of the expected doubling of Finnish electricity demand in the coming years.2151

Plans for Small Modular Reactors (SMRs)

From 2022 to 2024, a flurry of studies on SMRs were initiated and numerous MoUs signed (see section on Finland in WNISR2023 and WNISR2024). MoUs are not binding, and the timing of potential investment decisions has not been specified.

The most advanced project seems to be that of Finnish vendor Steady Energy, which signed a letter of intent with utility Helen in October 2023 to work towards a multipurpose agreement, ultimately enabling the commissioning of up to ten SMRs.2152 Helen launched a “nuclear energy program” in September 2024 to evaluate potential partners, suppliers, sites, and SMR designs for heat and power generation. The company stated that it hoped to operate its first SMR for district heating in Helsinki by the early 2030s.2153

In early 2024, nuclear operator Fortum launched a pre-licensing dialogue with the Finnish nuclear safety authority while reviewing the potential of several sites for new nuclear capacity, both “small and large”. However, Fortum Vice-President Laurent Leveugle pointed out that the company was “unlikely to embark on a first-of-a-kind facility.”2154 This statement clarified there was little, if any, expectation on Finnish SMR construction for commercial power production to begin anytime soon, especially since most companies with whom MoUs have been signed remain far from building their first reactors and none have any operational experience or even a certified SMR design (see chapter on SMRs).

In March 2025, the results of Fortum’s two-year feasibility study on new nuclear in Finland and Sweden—the company co-owns four Swedish reactors—were presented.2155 In a nutshell, Fortum concludes that for new nuclear projects, lead times of ten years or more are required and that the company will not be able to fund any projects itself; thus it would require some form of risk-sharing with respective states, and capital investment sharing with potential project partners, before an investment decision can be taken towards the end of the decade. In short, the company does not expect new nuclear to deliver any power in Finland before the second half of the 2030s.

Notably, Fortum concludes that for an effective project execution, they must select a “proven technology”, “[build] more than one reactor”, and “rely on competent owners with nuclear experience.” As a result, Fortum says it will “deepen collaboration” with three potential vendors: Westinghouse, EDF, and GE-Hitachi (GEH).2156 For Westinghouse and EDF, it remains uncertain whether the cited large reactor designs, i.e., AP-1000 and EPR, will be built again by the vendors, given disastrous experiences at past projects. Indeed, according to the Wall Street Journal, Westinghouse signaled in 2024 that it “no longer takes on reactor construction,”2157 while EDF is working to incorporate the lessons of said experiences to develop an improved (but yet unproven) design, the EPR2, and GEH’s BWRX-300 has not received a general design certification by any regulator yet, but did receive a construction license in Canada (see chapter on SMRs).

Energy Policy Outlook

According to global energy think tank EMBER, Finland produced 84.27 TWh (gross) of electricity in 2024. Nuclear power had the highest share at 43.6 percent,2158 followed by onshore wind (26.8 percent), hydro (19.1 percent), biomass (13.8 percent), coal (1.8 percent), and other fossil fuels (4.4 percent). Solar provided 1.6 percent of the power mix, while gas provided 0.9 percent.2159 The only Finnish offshore wind farm near Pori2160 provided 0.4 percent of the country’s electricity.2161

Finland is committed to fully decarbonize its electricity system by 2035.2162 Over the past decade, renewable capacities have grown substantially. Wind grew thirteenfold in the decade from 2014 to 2024, from 627 MW to 8.4 GW, driven mostly by onshore wind expansion. In 2024, solar PV capacity reached a modest 1.2 GW.2163

According to the Finnish National Energy and Climate Plan, submitted in June 2024, by 2035 wind power generation is to increase to over 46 TWh and solar generation to roughly 7 TWh from a negligible level today. For 2040, the plan envisions a nuclear share of around 30 percent, with absolute electricity generation increasing to around 37 TWh in 2025 and then remaining roughly stable. Wind power shall contribute approximately 40 percent by 2040 while Biomass and hydro shall account for 10 percent and 12 percent, respectively; the remainder is to be covered mainly through solar PV.2164

In August 2024, the Ministry for Economic Affairs and Employment published a 17-point action plan to promote the expansion of offshore wind power capacities. The report notes that there were inquiries for offshore grid connections adding up to 90 GW (albeit with a partial overlap), and that Fingrid identified five potential connection areas for wind farms off the Finnish coast with up to 1.3 GW grid capacity each.2165 In their latest update published in November 2024, Fingrid expanded their potential sites by two, bringing the potential offshore grid connection capacity to 9 GW in the 2030s.2166

France

See Focus Countries – France Focus.

The Netherlands

The Netherlands operates a single, over 50-year-old 482-MW PWR at Borssele—the oldest in the E.U.—that provided 3.4 TWh of electricity in 2024, a 10 percent decline over 2023. This corresponded to a stable 2.9 percent of the country’s electricity, compared to the historic maximum of 6.2 percent in 1986, when the country also operated a 60-MW BWR at Dodewaard. This reactor operated from 1968 to 1997. All the spent fuel had been removed by April 2003, and the site entered a 40-year “safe enclosure period” in June 2005, after which the plant is to be dismantled2167 (see Decommissioning Status Report). Refer to Netherlands Focus in WNISR2024 for a detailed description of the changes in the Dutch nuclear power policy in recent years.

Lifetime extension plans for the Borssele plant for up to twenty years beyond 2033, potentially allowing the reactor to operate for 80 years, seem to be advancing, but a final decision has yet to be made. The government’s multi-year 2026 Climate Fund Plan, released in April 2025, notes that the government will be “investigating the safe and efficient extension of the lifespan of the nuclear power plant in Borssele.”2168 In November 2024, the IAEA conducted a pre-SALTO (Safety Aspects of Long Term Operation) mission to the plant and indicated 15 issues that would have to be addressed to bring the plant up to “the level of IAEA safety standards and international best practices.” This includes the finalization of “aging management [programs] for mechanical and electrical components” as well as civil structures and ensuring that skilled personnel is available for long term operation.2169

The plant is operated by Elektriciteits-Produktiemaatschappij Zuid-Nederland (EPZ). Seventy percent of the plant’s shares are owned by Zeeuwse Energie Houdstermaatschappij (ZEH), which is in turn owned by the province of Zeeland and several municipalities in the same province. German utility RWE owns the remaining shares.2170 The lifetime extension requires amendments to the so-called Borssele covenant, an agreement that was introduced for the previous lifetime extension to 2033, as well as to the Nuclear Energy Act, among other legislative changes.2171

In an interview with news agency Omroep Zeeland, EPZ manager Tom Keij said that EPZ would need to know by 2029 whether Borssele shall operate beyond 2033, or whether it shall be decommissioned.2172 The interview was the subject of a written inquiry to the Dutch Parliament that was answered in May 2025 by Sophie Hermans, now-acting Climate and Energy Minister. The government hoped to pass necessary regulatory amendments by 2026 so that EPZ could submit its application to the regulatory agency in 2027.2173 However, in June 2025, the Schoof Government collapsed, and snap elections were called for 29 October 2025.2174 Additionally, the lifetime extension remains subject to a shareholder vote, expected in 2029. EPZ’s shareholders have indicated that they would only agree to the project if the state took over most of the financial risk.2175 Thus, it remains to be seen whether the new government will pursue the lifetime extension, and if so, whether Dutch regulators will approve.

In recent years, despite governmental advisory committee warnings of potential overcapacities in the grid,2176 Dutch energy policy has shifted towards nuclear newbuild (see Netherlands Focus in WNISR2024 for details). The previous government had suggested deciding on whether to pursue the construction of two new nuclear reactors in the Netherlands by the end of 2024.2177 After the 2023 election, in its coalition agreement (2024), the incoming right-wing government expanded the target to four new reactors—as did a motion passed by Parliament in February 2024—and pledged a total of €14 billion (US$15.2 billion) to the nuclear program until 2035.2178 These funds were set aside in the 2026 Climate Fund Plan, but have been deemed “likely insufficient” to build all four reactors.2179 Answering a written parliamentary question in February 2025, Energy Minister Sophie Hermans said that the plan to select a series of potentially suitable locations for newbuild, including the current Borssele site, by 2025 would not be possible and concluded that “it does not seem realistic anymore to have the first plant in operation by 2035.”2180 The choice of the Borssele site would likely require substantial grid expansion investments to cope in parallel with the high generation from nearby wind power,2181 and another potential site in the province of Groningen is facing local opposition from city councils and the province.2182

In many of the cited references, the Dutch words “kerncentrale” and “kernreactor” are used seemingly interchangeably, sometimes even in the same document.2183 They translate to “nuclear power plant” and “nuclear reactor”, respectively. There are substantial differences regarding cost as well as siting, licensing, and construction procedures between building four reactors in total and four power plants in different locations, with potentially more than one reactor per site. Until the government announces concrete capacity targets, WNISR considers the expansion plans to relate to four reactors. The committed government funding for the entire nuclear program is very limited whatever the final target definition would be.

In another progress report to the Tweede Kamer (House of Representatives), published in May 2025, Minister Hermans warned that the earmarked sum of €14 billion would not suffice (even for two reactors) and that €20–30 billion (US$22.7–34 billion) would be required to fund two reactors. She added that the Dutch government would have to finance the projects during the “first phases of construction”, and even then it would be challenging to attract private investors into the project.2184 In mid-2024, the Dutch pension fund floated the idea of investing in the projects.2185

Meanwhile, in November 2024, engineering firm Amentum was tasked by the Dutch Authority for Nuclear Safety and Radiation Protection (ANVS) to evaluate the potential of constructing two new PWRs by 2035 from three vendors, Westinghouse, KHNP, and EDF, based on feasibility studies carried out by the vendors.2186 In late March 2025, ANVS concluded from the Amentum report that, from a safety perspective, all three designs “seem suitable for construction in the Netherlands,” though the review “does not guarantee that we [ANVS] will be able to issue a license.” A detailed assessment of the design, depending on the chosen site, would be necessary before a definitive permit could be granted.2187 By mid-March 2025, KHNP had already announced it would no longer pursue the project of building a reactor in the Netherlands, leaving only Westinghouse and EDF as potential vendors.2188

In March 2025, a preliminary study by NRG Pallas, the operator of a medical isotope facility at Petten, on the integration of SMRs, commissioned by the Province of Gelderland, identified “four areas” as “potentially suitable for the construction of an SMR.”2189 Reportedly, representatives of the province said that they were going to “do follow-up research and enter into concrete discussions on the “arrival of SMRs.”2190 The ambition is to designate two locations by 2027 and to deploy a first unit between 2035–2040.2191

In October 2024, a joint report of NRG Pallas with the Dutch Agency for Applied Research TNO on SMR deployment in the Netherlands concluded that the construction of “two to more than 13 SMRs”, supplying both electricity and heat, would be viable by 2050.2192 The report also stated:

On the other hand, with delayed introduction of SMRs or no nuclear at all, a carbon neutral energy system in 2050 is possible as well (…).

Nevertheless, it can be concluded that SMRs are an important option for decarbonization of the industry by supplying process heat.

However, apart from several non-binding agreements (see previous WNISR editions), there have been no commitments to SMRs in the Netherlands, and the central administration seems to be focused on high-capacity reactors.

Today’s electricity mix in the Netherlands is dominated by natural gas, but the situation is changing fast. In 2024, renewables combined generated 50.9 percent of total power—up from 26.8 percent in 2020—that is more than all fossil fuels and nuclear combined at 49 percent, with wind energy contributing 27.3 percent, solar 18.1 percent, and bioenergy 5.5 percent compared to natural gas providing 35.8 percent, coal 7 percent, “other fossil fuels” 3.3 percent, and nuclear 2.9 percent.2193

Renewable capacities continue to grow, with over 200 MW of onshore and 770 MW of offshore wind capacities added in 2024, in addition to a remarkable 2.4 GW of new solar.2194 According to REN21, the Netherlands has the highest installed photovoltaics capacity per capita in the world.2195 Nevertheless, in a November 2024 report, the Environmental Agency noted that additional policy changes were necessary to ensure that the 2030 emission reduction targets were met.2196

Spain

As of 1 July 2025, Spain operates seven reactors totaling 7.1 GW capacity that provided 51.9 TWh in 2024, representing a 4 percent decrease year-on-year and 19.1 percent of the country’s electricity generation—far below the historic maximum of 38.2 percent in 1989. Spain’s reactors have a mean operating age of 40.4 years as of mid-2025.

Spanish nuclear ownership is concentrated with the utilities Iberdrola and Endesa. Both utilities share ownership of Ascó-2 and Vandellós-2, Almaraz-1 and -2 with Naturgy, and Trillo with Naturgy and EDP. Endesa is the sole owner of Ascó-1, and Iberdrola fully owns Cofrentes.2197

In January 2019, Spain’s coalition government agreed on a nuclear phaseout plan with utilities Endesa, Iberdrola, and Naturgy as part of the overall Integrated National Energy and Climate Plan (INECP).2198 All of Spain’s reactors are scheduled to be closed by 2035. However, the policy also secured the possibility for all reactors to apply for lifetime extensions beyond 40 years, in contrast to the previously governing Socialist Party’s (PSOE) policy allowing for a limited seven-year deferral of the plants’ closures initially scheduled for 2024–2028.2199 This plan was confirmed in the final version of the country’s NECP, published in September 2024.2200

Operating licenses were extended beyond 40 years for all the Spanish reactors within the past five years (see previous WNISR editions). As shown in Table 22, Almaraz-1 and -2 are the next reactors to close in November 2027 and October 2028, respectively. As the current operating licenses for Ascó-2, Vandellós-2, and Trillo expire prior to their planned delayed closure, they will have to be extended if the reactors are to operate according to the latest schedule.

1. Status of the Spanish Nuclear Fleet (as of 1 July 2025)

Reactor	Net Capacity (MW)(a)	Grid Connection	License Expiration(b)	Expected Closure(c)
Almaraz-1	1 011	01/05/1981	November 2027	November 2027
Almaraz-2	1 006	08/10/1983	October 2028	October 2028
Ascó-1	995	13/08/1983	October 2030	October 2030
Cofrentes	1 064	14/10/1984	November 2030	November 2030
Ascó-2	997	23/10/1985	October 2031	September 2032
Vandellós-2	1 047	12/12/1987	July 2030	February 2035
Trillo-1	1 003	23/05/1988	November 2034	May 2035

Sources: Various, compiled by WNISR, 2025

Notes:

a - Reference Unit Power from IAEA-PRIS as of 1 July 2025.

b - Individual reactor pages on Spanish Nuclear Safety Council’s website.2201

c - Nuclear Safety Council, Eighth and Ninth Report to the Convention on Nuclear Safety.2202

In recent years, the Spanish electricity market has seen an increase in both the absolute number and the duration of periods with very low or even negative electricity prices.2203 For example, during the Easter holidays in March 2024, the temporary reduction of nuclear capacities to only 3.6 GW was attributed partially to a “prolonged renewable output,” during which Spanish wholesale electricity prices were at less than €0.6 per MWh (US$20240.65/MWh).2204 The pattern repeated in 2025, when Almaraz-1 and -2, Ascó-1, and Cofrentes were disconnected from the grid due to low electricity prices induced by increased supply from wind and hydro, pushing wholesale prices to less than €12.14 per MWh (US$13.8/MWh) in early April 2025.2205

Spanish nuclear operators and lobby group Foro Nuclear blame a “disproportionate tax burden”, which increases operating costs and forces the plants to shut down temporarily for economic reasons.2206 Indeed, the levy for nuclear waste management and decommissioning activities was raised from €7.98/MWh to €10.36/MWh (US$9.1 to US$11.8/MWh) in June 2024 due to “rising storage and disposal costs”, estimated at €20.2 billion (US$202321.8 billion) in the latest General Plan for Nuclear Waste Management approved in December 2023. This tax increase is subject to a legal challenge submitted to the Supreme Court by Foro Nuclear, which has been lamenting the tax burden weighing on nuclear operators for years (see past WNISR editions).2207 Iberdrola and Endesa have both filed appeals contesting the General Plan and the derived fee revision, arguing it constitutes a breach of the 2019 phaseout protocol, and claim a combined €778 million (US$883.5 million) in compensation.2208

Attempts to Reverse Spanish Nuclear Phaseout Policy

The Spanish plans to end operations of their nuclear power plants have been facing increasing opposition in recent years (refer to previous WNISR editions). The 2023 elections led to conservative party, Partido Popular (PP), winning the majority of seats in the Spanish Parliament. However, after they failed to form a government, incumbent socialist Prime Minister Pedro Sánchez formed a government, which remains in power as of mid-2025, despite mounting pressure due to ongoing investigations into corruption allegations against senior officials in his government and party, as well as family members.2209

The PP had run its election campaign on reversing the nuclear phaseout plan,2210 and in February 2025, it passed a non-binding law proposal in the lower house of the Spanish Parliament to “extend the useful life of existing nuclear power plants in our country” and to

facilitate the economic sustainability of nuclear facilities in a way that reflects the fundamental role of this technology in ensuring a secure and stable electricity supply, and in its contribution to reducing electricity market prices and greenhouse gas emissions.2211

The vote showed the divide among Spanish political parties, with 171 votes in favor from PP, far-right party Vox, and the Navarrese People’s Union (Unión del Pueblo Navarro or UPN), against 164 votes from left-wing and supporting parties, and 14 abstentions from Catalan separatist parties.2212

Industry demands to extend the reactors’ lifetimes beyond 2035, which had been ongoing for several years, were intensified after the vote.2213 A few days after the parliamentary vote, six companies from the nuclear industry—with the support of 35 companies from other sectors—signed a manifesto calling for “dialogue and renegotiation of the 2019 agreement” and a revision to “the National Integrated Energy and Climate Plan to incorporate measures ensuring the continuity of nuclear energy,” among other demands.2214

Also in February 2025, and regularly thereafter, a number of organizations and prominent stakeholders called for the extension of operating lifetimes. These included María Guardiola, president of Extremadura (the region bordering Portugal in which the Almaraz plant is located), who also suggested that the Spanish state should take over the operations of the plant if the taxes were not reduced.2215 Co-signing a declaration with María Guardiola to reverse the Almaraz closure decision, the president of the Community of Madrid, Isabel Díaz Ayuso, described the nuclear plant as “essential” to guarantee the stability of the national electricity system and regional economic development.2216

Additionally, in March 2025, Iberdrola’s president, Ignacio Sánchez Galán, claimed that the phaseout would increase Spanish wholesale electricity prices by up to 30 percent and have a detrimental impact on grid reliability. He further questioned, “Can we as Europeans be in a position to renounce those natural energy resources, just because of ideology? Or do we have to be pragmatic, like the Americans?”, referring to lifetime extensions and recommissioning of closed reactors in the U.S.,2217 and affirmed that nuclear was “absolutely necessary to keep the lights on”.2218

Effects of the Iberian Blackout on 28 April 2025

On 28 April 2025, the Iberian Peninsula experienced the worst power blackout in its recent history, affecting mainland Spain and Portugal. Minor power cuts were also experienced in French border towns and in Andorra.2219 Power was lost shortly after 12:30 a.m. and restored throughout the next day with help from neighboring grid connections in France and Morocco, as well as hydro-electric and pumped storage plants and combined-cycle gas plants.2220

Despite the Spanish government asking for patience until the cause for the blackout was identified, and grid operator Red Eléctrica narrowing the potential causes down to two substations that had failed in the southwest of Spain, the high share of renewables in the electricity mix at the time of the blackout—solar and wind were providing about 70 percent of power—fueled calls for the reversal of the phaseout policy.2221 The search for the cause of the blackout began shortly after the incident, and cyberattacks were quickly ruled out. Despite this uncertainty, the high share of renewables in power production continued to be blamed, with commentators confusing correlation with causation.2222

On 17 June 2025, the Spanish government reported the conclusion of the investigation carried out by a specially tasked committee.2223 A full report with individual retracted sections was published on the same day.2224 According to the analysis, a series of individual events triggered a chain reaction within the system.2225 These events include power plants capable of stabilizing the grid not behaving as had been planned by the grid operator, an unexpectedly slow response of these plants after being called to stabilize the grid when renewable production steadily increased on the morning of the 28 April, and so-called “rhythmic oscillations” as well as “premature disconnections” from the grid that led to a loss of capacity that was too fast for the grid to keep up with.2226

The Minister for the Ecological Transition, Sara Aagesen, concluded in her presentation of the report that grid-operator Red Eléctrica had “miscalculated the power capacity needs for that day,” failing to bring additional thermal capacity to the grid, and she blamed private operators who “were supposed to control voltage—[and] were paid to do just that—did not absorb all the voltage they were supposed to.” The companies were not named.2227 The report does acknowledge that “vulnerabilities and deficiencies” exist in the Spanish and Portuguese grids and advocates for improved supervision of operators, an increase in the country’s overall power capacity, and the expansion of interconnections to neighboring countries.

On 24 June 2025, only one week after the publication of the inquiry report, the Spanish government adopted a set of “urgent measures” to strengthen the electricity system in three areas:

• strengthening the monitoring, verification, compliance, and transfer capacities of the system
• enhancing the storage and flexibility of the system
• advancing electrification, facilitating the integration of industrial demand, transport, and renewable generation2228

One day prior to the release of the task force report, the European Investment Bank (EIB) announced funding of €1.6 billion (US$1.8 billion) to increase the power exchange capacity between Spain and France from 2.8 GW to 5 GW,2229 and in May 2025, the Spanish government launched a €700-million (US$795 million) support scheme to fund up to 3.5 GW for “innovative” energy storage projects.2230 Obviously, both projects had been in the pipeline for several years.

Meanwhile, the Spanish government remains committed to its target of becoming climate neutral by 2050 and reaching 81 percent of renewable electricity supply by 2030.2231 In 2024, 7.3 GW of renewable capacity was added to the grid, of which 6 GW were solar PV and 1.3 GW were onshore wind. Over the past decade, renewable capacities grew by 37 GW to over 85 GW in 2024 consisting of solar, wind, hydro, as well as marginal capacities of bioenergy and other technologies, such as geothermal.2232 The electricity mix (gross) in 2024 comprised 22.4 percent onshore wind, 20.9 percent solar, 19.6 percent nuclear, 18.7 percent natural gas, 11.6 percent hydro, followed by “other fossil” (3.6 percent), bioenergy (2.2 percent), and coal (0.9 percent).2233

Sweden

Sweden’s nuclear fleet of six reactors generated 48.1 TWh in 2024, a 4.4 percent increase compared to the previous year. Nuclear accounted for 29.4 percent of the country’s total electricity production. The share of nuclear power in the country’s electricity mix peaked in 1996 at 52.8 percent when 12 reactors were operating, while the fleet reached its highest output of over 75 TWh in 2004 with 11 units still on the grid. Refer to Sweden Focus in WNISR2024 for details on the Swedish nuclear energy policy and past developments.

Lifetime Extensions for Aging Swedish Fleet?

The mean age of the Swedish fleet is 43 years, with four reactors over 41 years (see Figure 87).

1. Age Distribution of the Swedish Nuclear Fleet

Sources: WNISR with IAEA-PRIS, 2025

In June 2024, Vattenfall, the majority-owner-operator of the three Boiling Water Reactors (BWRs) at Forsmark and the two Pressurized Water Reactors (PWRs) at Ringhals (that were connected to the grid between 1980 and 1985), announced that it had “taken a directional decision to extend the operating lifetime of the plants’ reactors from 60 to 80 years,” and therefore was awaiting an “in-depth investigation”.2234 This would push their closure dates into the 2060s. The investigation is planned to conclude in 2026, after which an investment decision is expected.2235 In September, the owners of Oskarshamn-3, Uniper and Fortum, declared the launch of a study to investigate the reactor’s operation beyond its current expected lifetime, 2045, “well into the mid 2060s.”2236

Independent of the outcome of the investigations and the decisions in principle to extend operating lifetimes, all of the reactors have to pass decennial safety reviews by the regulator.

In recent years, the operations of Swedish nuclear power reactors have been plagued by a series of unplanned and extended outages (see past WNISR editions). While the operations at the Ringhals site appear to be under control following major challenges in 2022,2237 Forsmark-3 was shut down on 1 September 2024 for a revision that was scheduled to conclude by mid-October of the same year.2238 In November, it was announced that the outage would be extended until 27 January 2025 due to extensive repairs on cracks affecting two low-pressure turbines, but that the annual closure planned for March 2025 would be combined with the audit scheduled for 2026, reducing the total planned outage of Forsmark-3 by 15 days for 2025 and 2026 combined.2239 The plant came back online during the first weekend of February 2025.2240 Only a few days later, on 4 February, the reactor went offline again due to a faulty valve and returned to the grid three days later.2241

On 28 March 2025, after achieving its longest run of consecutive days generating power per OKG, Oskarshamn-3 was taken off the grid for refueling and maintenance planned to last “just under three weeks.”2242 In April 2025, cracks were discovered in several critical pipes, leading to an extension of the outage until 15 June. By May, this had been further extended until mid-August,2243 and on 26 June restart was again delayed, to early September 2025 this time.2244

Swedish Nuclear Policy Reversal—Embracing Newbuild

In changing its target definition for electricity production by 2040 from “100% renewable” to “100% fossil-free” in June 2023, Sweden paved the way for nuclear newbuild in the country.2245 Current plans envision 2.5 GW of new nuclear on the grid by 2035 and the equivalent of ten new large reactors by 2045.2246

In January 2024, a “nuclear power coordinator” was appointed, tasked with identifying necessary measures, removing regulatory hurdles, coordinating with all stakeholders, and promoting new nuclear development. The final report is scheduled for end-2026.2247 In the second interim report, the coordinator lists several completed, ongoing, and planned promotional measures to support Swedish nuclear newbuild, for example, the potential financing of pilot and demonstration plants by the Swedish Energy Agency or the development of a financing model for new nuclear plants by the Ministry of Finance.

The coordinator notes that from a regulatory perspective, the acceleration of the licensing process for up to 2.5 GW of new nuclear is necessary and is “the most challenging aspect”. His report elaborates that the newly proposed permitting process is estimated to take 10 years from application to grid connection but acknowledges that “there are some uncertainties about the efficiency of the process when it is used for the first few times.” The report concludes, therefore, that parallel expansion projects involving “at least two sites” is the strategy most likely to achieve the targeted 2.5 GW deployment by 2035.2248

In its 2025 budget proposal adopted in late 2024, the government earmarked SEK1 billion (US$104 million) to fund fossil-free electricity generation. This included SEK100 million (US$10.4 million) to “support pilot and demonstration projects in the nuclear field”, SEK2.5 million (US$261,000) in increased funding to the Swedish Environmental Protection Agency for consultancy on developing new nuclear permitting processes, and SEK30 million (US$3.1 million) for further activities towards creating favorable conditions for new nuclear.2249

A report by the Swedish Energy Agency highlights another looming challenge: the potential future lack of competent staff, not limited to nuclear but also affecting the electrification of Sweden’s entire energy supply system. For nuclear, the report points out that, in the short-term, current plant staff members are nearing retirement age, while there is growing demand for expertise in project planning, construction, and operation of a potential new fleet. It notes that other countries likely experiencing a growing need for skilled staff at the same time, due to their own nuclear expansion plans and/or maintenance requirements, will cause fierce competition in the future and thus pose a major challenge needing to be addressed. Consequently, to account for necessary training and education times, the report calls for an assessment of future demand and supply for a nuclear-trained workforce.2250

Regarding the planned newbuild program’s financing, a commission tasked with developing models through which the government could help shoulder the associated risks concluded its work in August 2024.2251 A financing scheme based on the commission’s recommendations was proposed to the Riksdag, Sweden’s parliament, by the government in March 2025. The corresponding legislation was passed in May and will enter into force in August 2025.2252

Recognizing the private sector’s reluctance to take on the risk of nuclear newbuild, the new legislation allows the Swedish state to fund up to 75 percent of a project’s investment cost with government loans, which would correspond to SEK315 billion (US$32.9 billion) for four reactors.2253 The government notes that concrete details on funding and financing would be determined once a specific project’s costs have been determined—as of now, there is no project. Furthermore, the operator would receive fixed revenues via a symmetrical Contracts-for-Difference scheme capped at SEK0.8 per kWh (US$0.083/kWh). This is to be funded by a levy on electricity prices to electricity consumers of SEK0.02 per kWh (US$0.002/kWh).2254 Risks of cost increases are to be shared equally with private investors according to the share structure of the project. The scheme is still subject to approval by the European Commission.2255

In late June 2025, the government approved a new ordinance regulating the application process for state aid to companies seeking to build reactors. On 1 August 2025, the applications portal opened for “companies whose sole activity is the construction and operation of new nuclear power reactors.”2256

In an interview with Swedish newspaper Dagens Nyheter, Prime Minister Ulf Kristersson stated that the construction of a new nuclear power plant would commence before the next election, which is scheduled for September 2026.2257 This ambitious plan is described as unrealistic from both sides of the political spectrum, e.g., by former advisor and CEO of conservative think tank Timbro, P. M. Nilsson2258 and by Rolf Lindahl of Greenpeace.2259

The skepticism appears reasonable given that none of the Swedish operators have yet committed themselves to specific projects or vendors. In June 2024, Vattenfall’s head of new nuclear power, Desirée Comstedt, said that they “[had not] made a choice of reactor technology yet” but that two “potential suppliers of small modular reactors” had been shortlisted. These were Rolls-Royce and GE Hitachi. However, she noted that “Vattenfall will continue to investigate the conditions for building large-scale reactors.” Potential vendors were Westinghouse, EDF, and KHNP.2260

Thus, Vattenfall has not even decided whether to build high-capacity plants or wait for currently unavailable SMR technologies, but made clear it wants to build at a site adjacent to Ringhals.2261 It noted in a report from 2024 that it would wait to make an investment decision until a final design of a reactor has been approved (see chapter on SMRs) and not subject to further modifications.2262 Nevertheless, in April 2025, the company announced it was in the process of setting up Videberg Kraft AB, a project company exclusively for nuclear, in order to be able to apply for the new state aid mechanism.2263 By that time, only four technology suppliers were still under consideration.

Finnish utility Fortum owns shares of the Oskarshamn-3 reactor. The utility has announced ambitious plans to build new nuclear power plants in Finland and Sweden (see also section on Finland). In March 2025, the company concluded a two-year feasibility study clarifying that it maintained its interest in new nuclear but did not view it as a viable option in the short term, seeing an opportunity for deployment in the second half of the 2030s at the earliest “if market and regulatory conditions are right.” Its president and CEO commented that “In the next 5-10 years, new demand in the Nordics will be primarily met with new onshore wind and solar power coupled with flexibility and storage solutions as well as lifetime extensions of existing nuclear power plants.” The company said, however, it would now “continue to deepen the collaboration with two conventional reactor technology providers, EDF and Westinghouse-Hyundai, and one SMR developer, GE-Hitachi,” the stated intent being to “mitigate project risks before we would consider an investment.”2264 In June 2025, Fortum reportedly signed an Early Work Agreement with EDF to “continue their collaboration toward the potential development of new nuclear projects in Finland and Sweden.”2265 The exact meaning of this remains unclear.

The Swedish government’s nuclear power plant expansion plans are criticized as “energy populism” by Greenpeace campaigner Rolf Lindahl. According to his analysis, the shift from an ambitious (and successful) renewables expansion plan to nuclear might risk Swedish climate targets because the plants will probably not be built—or they might be delayed—and renewables are being neglected.2266 The headwinds for the expansion of renewables is evidenced by only one of sixteen planned onshore wind projects moving along in the permission process in 2024,2267 or the cancellation in late 2024 of 13 offshore wind projects (totaling ~32 GW) because of “unacceptable consequences for Sweden’s military defense.” This, however, might be an effect of Sweden’s particular permit acquisition scheme.2268 Subsidies for solar power and tax credits for small producers were lowered or scrapped altogether in 2025, with the government reportedly arguing that the sharp increase in installations demonstrated that there was no longer a need to subsidize solar power. The country saw a record increase of PV capacities of 1.6 GW, including small and large facilities, in 2023.2269 Another 1 GW of PV, added in 2024, increased the total installed capacity to approximately 5 GW.2270

Consequentially, in 2024, 69.5 percent of Swedish gross electricity production, totaling 173.3 TWh, was covered by renewables, i.e., hydro (37.6 percent), wind (23.2 percent), bioenergy (6 percent), solar PV (2.1 percent), and offshore wind (0.3 percent). Nuclear power supplied 29.2 percent, while the remainder was covered by “other fossil” fuels (1.3 percent) and gas (0.1 percent).2271

Switzerland

Swiss gross nuclear power production declined slightly by 1.5 percent year-on-year to 23 TWh in 2024. This corresponded to 28.4 percent of gross national power production, which reached a record level of 81 TWh, up 11.4 percent compared to 2023, fueled mainly by a substantial increase in hydro output.2272 The historic maximum nuclear share of 43 percent was achieved in 1996.

With an average reactor age of 49.3 years (see Figure 88), Switzerland operates the second oldest nuclear fleet in the world (behind the Netherlands that operates only one 52-year old unit), of which Beznau-1, age 56, is the oldest commercially operating reactor in the world. Beznau-2 is almost 54 years old. The reactor at Mühleberg was closed in 2019 and is being decommissioned (see Switzerland in Decommissioning Status Report).

1. Age Distribution of the Swiss Nuclear Fleet

Sources: WNISR with IAEA-PRIS, 2025

Lifetime Extensions

Swiss nuclear licenses are granted for an indefinite period. They can be revoked if the sites are deemed unsafe by the regulator.2273 Currently, the operators of all still-running power plants are seeking to operate the plants for at least 60 years,2274 despite controversial reports regarding the operating safety at Beznau-12275 and Leibstadt2276 (see previous editions of the WNISR for details).

A study on “the technical feasibility of operating the Beznau nuclear power plant beyond 60 years” was launched in March 2024.2277 In December 2024, the operator of the plant, Axpo, announced that Unit 1 would close in 2033 and Unit 2 in 2032, which would extend their operations to 64 and 61 years, respectively. Axpo would invest an additional CHF350 million (US2024$398 million) for safety upgrades after already having invested CHF2.5 billion (US$20242.8 billion) since the commissioning of the plant.2278

In December 2024, the Swiss Federal Nuclear Safety Inspectorate (ENSI) stated that it took note of the decision by the operator Axpo to operate Beznau up to 2033. The next Periodic Safety Review (PSR) at Beznau is scheduled for 2027, and Axpo must officially submit the corresponding documents to ENSI before the end of 2027.2279 But, according to a knowledgeable source, due to the fact that the decommissioning application must be submitted in 2028 to close the plant in 2033, the documents for the PSR would have to be submitted earlier than 2027.

While conservative National Council member Christian Imark claims that the plant could operate for 100 years, Andreas Pautz of the ETH Lausanne and the Paul-Scherrer Institute says that possibilities to extend operations were limited due to flaws in the steel of the reactor pressure vessel of Beznau-1.2280 These were first discovered in 2015.2281

Furthermore, Beznau’s ability to provide a reliable electricity supply for the next few years remains questionable, as the frequency of unplanned outages has reportedly been increasing at the site.2282 For example, in August 2024, Beznau-1 was taken off the grid for a few hours after a valve failure in the “non-nuclear” part of the site.2283 It was offline again from 28 September to 28 October at first for repairs to a power cable and then due to steel sample results from feedwater tanks requiring “further clarification”, both issues in non-nuclear areas but still impacting operation.2284 According to IAEA-PRIS, Beznau-1’s load factor plunged from 92 percent in 2023 to 76 percent in 2024.

Further, during the summers of 2024 and 2025, output at the plant had to be reduced on several occasions due to high temperature levels of the Aare River, e.g. in early July 2025, after operating at half capacity, both units were eventually temporarily shut down.2285

As of July 2024, operators of Gösgen and Leibstadt had not extensively investigated the possibility of operating the two reactors past 60 years, the main holdback being, according to them, the uncertainty over political acceptance.2286 Nevertheless, the companies confirmed at the time that they had been looking into the implications of extended operation—as was known for some time—and that, for planning purposes, an extension would have to be notified 15 years prior to the latest announced closure date. As of 2025, plans still provide for the closure of Gösgen in 2039 and Leibstadt in 2044, corresponding to 60 years of operations.2287 While there have been no official updates, some political actors are already advocating for 80 years of operations.2288

Political Nuclear Newbuild Initiatives

In 2018, the Swiss Energy Strategy 2050 entered into force after passing by referendum in 2017. With this came the ban on building new nuclear power plants in Switzerland.2289 Since the Russian invasion of Ukraine, concerns over energy supply security have resurfaced, and with these, calls for lifting the ban on nuclear newbuild have intensified, too. This culminated in the submission of an initiative (so-called Volksinitiative or Initiative Populaire that can be brought to referendum by citizens) called “Electricity for all at all times (stop the blackout)” that had gathered the necessary number of at least 100,000 signatures in February 2024.2290

Meanwhile, the rapid expansion of domestic renewable capacities was approved by referendum in June 2024.2291 Refer to previous WNISR editions for more details on Swiss Energy Policy.

In August 2024, the Federal Council formally declined the “Stop the Blackout” initiative to hold a referendum, but signaled its intent to issue a counter-proposal instead with legislative steps designed to deliver on the main demands of the initiative.2292 Consultation on the government’s legislative proposal to repeal the ban on nuclear newbuild from the Nuclear Energy Act ran from December 2024 to April 2025.2293 The amendments, which are yet to be validated by parliament, gathered support from center-right parties and opposition from Greens and other left-leaning parties.2294 It remains uncertain whether this legislative change will pass, given a stark divide mainly along party lines and between Swiss local (canton) governments.2295 Throughout these efforts, the accelerated deployment of renewable energy sources was still presented as the highest priority. Also, a recent analysis by the Energy Commission of the Swiss Academies of Arts and Sciences concluded that “even in the case of the revocation of the newbuild prohibition, the commissioning of a new nuclear power plant is unlikely before around 2050.”2296

Meanwhile, Swiss electricity production is dominated by hydro, which supplied 48.4 TWh or 59.6 percent of the country’s electricity in 2024, up from 40.8 TWh or 56.1 percent in 2023. Nuclear power supplied 23 TWh in 2024, which corresponded to 28.4 percent. Solar PV output increased from 4.9 TWh (6.8 percent) in 2023 to 6 TWh (7.3 percent) in 2024,2297 with record capacity increases of 1.8 GW in 2024, resulting in over 8 GW of total installed solar capacity by December 2024; the pace of deployment is expected to slightly drop in 2025 and 2026—to 1.5–1.6 GW annual additions—before the impact of new regulations kicking in as of 2027 are expected to reaccelerate buildout.2298 The share of wind power in the Swiss electricity mix is marginal (0.2 percent). As of 2024, only 102 MW of onshore wind energy were connected to the grid.2299 Fossil fuels represented 1.7 percent of the total generation.2300

United Kingdom

See Focus Countries – United Kingdom Focus.

Central and Eastern Europe

Bulgaria

Bulgaria, like other Eastern European countries, inherited its nuclear reactors from the Soviet Union. The country began its nuclear energy program in the 1970s with the construction of VVER-440 reactors. By the late 1980s, Bulgaria had also successfully completed two VVER-1000 reactors. A project to build four additional VVER-1000 reactors at the Belene site was initiated but ultimately abandoned following the collapse of the Soviet Union in 1991.

Bulgaria’s two operating reactors at the Kozloduy nuclear power plant generated 15 TWh of electricity in 2024, a slight decrease from the 15.4 TWh produced in 2023.2301 Despite this reduction in electricity output, the share of electricity generated from nuclear power increased from 40.3 percent in 2023 to 41.8 percent in 2024. This was due to a significant decline in electricity production from coal.2302

The Bulgarian energy sector has historically been heavily dependent on Russian imports, including both fossil fuels and nuclear fuel. However, following Russia’s invasion of Ukraine, this dependence has been significantly reduced. Prior to the war, over 90 percent of Bulgaria’s natural gas supply came from Russia, but this supply has been completely halted by Russia.2303

Since 2023, Bulgaria has met its natural gas demand through liquefied natural gas (LNG) terminals in Greece and Türkiye. Additionally, Bulgaria has signed long-term agreements with the Turkish company BOTAŞ, securing access to its LNG terminals and pipeline infrastructure.2304

Despite terminating direct imports of Russian natural gas for its own consumption, Bulgaria continues to serve as a gateway and transit country for Russian gas to Europe via the Balkan Stream pipeline, gas supplies through which reached a historic high in 2024.2305 Operational since January 2021, Balkan Stream is an extension of the TurkStream project and serves as a key conduit for Russian gas exports to southeastern and central Europe, including countries such as Hungary, Serbia, and Romania. However, the project has brought Bulgaria minimal financial benefits, as transit tariffs are heavily subsidized, and the bulk of profits are captured by Russia’s Gazprom and intermediaries in other countries.2306

Fuel Supply Diversification

Bulgaria remains reliant on Russian equipment and fuel for its Kozloduy nuclear power plant but is actively attempting to diversify nuclear fuel supply. In November 2022, Bulgaria’s Parliament voted to reduce dependence on Russian sources,2307 despite an existing contract with Russia’s TVEL for fuel provision until 2025 (included).2308 To align with the E.U.’s energy security strategy, Kozloduy’s operator signed a ten-year contract with Westinghouse in December 2022 to supply fuel for Unit 5 starting in 2024.2309 The first Westinghouse fuel arrived in April 2024.2310 The following month Unit 5 was partially refueled with 43 Westinghouse-manufactured assemblies,2311 and in May 2025 another 42 assemblies were loaded.2312 The phased transition to alternative fuel is to be completed by 2027. Additionally, in December 2024, Kozloduy signed a contract for a safety analysis to license the Westinghouse fuel assembly design for Unit 6.2313

Besides the Westinghouse deal, the Kozloduy operator also signed a 10-year supply contract (2025–2034) with Framatome for Unit 6.2314

Lifetime Extensions

The Kozloduy plant originally hosted six reactors, but the oldest four (VVER-440/v-230) were to be closed under a 1992 G7 agreement, as it was determined that they could not be economically upgraded to meet safety standards. Their closure was implemented in 2002 and 2006 as part of Bulgaria’s E.U. accession process which concluded in 2007.2315 The two operational VVER-1000 reactors (Units 5 and 6), commissioned in 1987 and 1991, are undergoing a refurbishment program to extend their lifetimes to 60 years, compared to the original 30-year licenses.2316 Regulations originally limited renewed operational licenses to 10 years,2317 so Unit 5 received a new license in 2017 to operate until 2027, and Unit 6 was relicensed in 2019 to operate until 2029.2318

However, following amendments to the Act on the Safe Use of Nuclear Energy, licenses for both units were revised in mid-2024 and are now indefinite.2319 Periodic safety reviews are still to be carried out at least every ten years, the results of which are to be approved by the regulator who specifies that “in the event that the results submitted do not show the necessary compliance with the safety requirements, the Chairman shall reasonably refuse to issue an order for approval, which shall constitute grounds for revoking the license concerned.”2320

Newbuild Projects

Efforts to expand nuclear capacity on an “accelerated schedule” continued in recent years despite the volatile political situation (see past WNISR editions for further details and earlier developments). In March 2023, state-owned Kozloduy NPP-Newbuild signed an MoU with vendor Westinghouse to plan one or two AP-1000 reactors at the site.2321 The MoU was formalized in June 2023 through a Front-End Engineering and Design (FEED) contract,2322 which was extended in October 2024.2323

An invitation for expressions of interest in the “engineering, construction, procurement and commissioning of new nuclear power plant on the approved site in Kozloduy with AP1000® technology” was issued in January 2024,2324 and by February, five companies—“Fluor BV, Bechtel Nuclear Power Company Limited, Hyundai Engineering & Construction Co., Ltd., a consortium led by China National Nuclear Corporation Overseas Ltd. and partner China Energy Engineering Group Tianjin Electric Power Construction Co., Ltd., as well as China Energy Engineering Corporation Limited”—had expressed interest.2325 Earlier press reports had also identified EDF (France) as a candidate.2326 Russian bidders had been barred from the process.

The same month, only Hyundai was preselected,2327 and the three parties—Westinghouse Electric Company, Hyundai Engineering & Construction, and Kozloduy NPP-Newbuild—finally signed an Engineering Services Contract in November 2024, making the project “irreversible” per then-caretaker Energy Minister Vladimir Malinov.2328 The contract runs for 12 months, and it covers site planning, licensing, permitting, project management, and operations & maintenance development for two AP-1000 reactors at Kozloduy.

As reported in WNISR2024, negotiations pertaining to the project’s financing also began in 2024. Since then, discussions with the U.S. Export-Import (EXIM) Bank have led to the issuance of a US$8.6 billion Letter of Interest.2329 The U.S. EXIM Bank states in no ambiguous words:

EXIM is prepared to support American energy dominance, consistent with President Trump’s agenda by financing nuclear energy investments in Bulgaria which support American businesses, create jobs and prosperity for families and workers in America, and reduce the global influence of malign and adversarial states.2330

The Export-Import Bank of Korea is also expected to formulate an offer; it had vowed to submit such a document by September 2024, though it does not appear to have followed through as of mid-2025.2331 Time is running out to meet the deadline set by the Bulgarian Parliament.

Indeed, based on a timeline decided by Parliament in December 2023, financing was to be secured by 30 July 2025, while the Final Investment Decision (FID) should occur by 31 August 2025 and be immediately followed by the EPC contract.2332 In February 2025, the Bulgarian Minister of Energy confirmed that “a final investment decision is to be made by the end of the year and by then the financing will be secured and only then will the EPC contract be signed.”2333

As for the project’s costs, then-Bulgarian Energy Minister Ruman Radev had stated in February 2024 that they “should not exceed 14 billion” (he did not specify the currency at the time).2334 Reportedly, he further said that the cost of electricity generated by the new units would be capped at €65/MWh (US$70/MWh), and that the indicative price was to be announced in a report by Westinghouse, due in March 2024.2335 As of mid-2025, WNISR has no knowledge of details from said report that would have been made available to the public and is not aware of any official estimate—be it provisionary—having been released. In January 2025, Patrick Fragman, then-CEO of Westinghouse said in an interview “We should have much more narrow estimates about the project cost around the end of 2025,” or in other words “in 2026 we will have a cost estimate for the new units of Kozoloduy NPP.”2336

The parliamentary timeline provides for the first unit to be commissioned by 31 December 2034.2337 Westinghouse has also indicated that commercial operation would begin ten years from now, in 2035.

The Belene Saga

Construction of a nuclear power plant at Belene in northern Bulgaria began in 1987 but was halted in 1990 and suspended in 1991. Construction resumed in 2008 but was abandoned again in 2012.

In October 2023, the Bulgarian government officially terminated the latest Belene nuclear project.2338 A few months prior, Parliament had authorized negotiations to potentially sell key equipment from Belene to Ukraine for the planned resumption of construction of Units 3 and 4 at the Khmelnytskyi nuclear plant in western Ukraine.2339 Political uncertainty over a joint E.U. stance on financial aid to Ukraine briefly disrupted proceedings in January 2024.2340 Nevertheless, in May 2024, a Ukrainian delegation, joined by U.S. nuclear experts, inspected the equipment at the Belene site.2341

Although the Bulgarian Parliament decided in late 2023 to allocate funds from the equipment’s sale to the Kozloduy construction project,2342 completed the verification of the suitability and storage of the equipment in June 2024,2343 and Parliament extended the negotiation deadline by 180 days in September 2024,2344 the sale—approved by the Ukrainian parliament in 20252345—was canceled by Bulgaria’s minority government in April 2025.2346 However, Bulgarian Prime Minister Rossen Jeliazkov commented that only the National Assembly could decide on the matter,2347 and in June 2025 then-Ukrainian Energy Minister German Galushchenko was quoted in Interfax-Ukraine as stating that “the issue is not closed, we have no official refusal from Bulgaria to continue negotiations.”2348

WNISR2024 had reported that “the idea [to complete Belene] was definitely abandoned in 2023,” but it seems new ideas have emerged. Voices in Bulgarian government and some influential politicians now suggest building AI data centers in partnership with the U.S., powered by nuclear energy from Belene, hinting at a potential revival of the Belene project.2349 However, the hypothetic project’s future remains entirely speculative.2350 In the end the “decision” not to sell the equipment to Ukraine raises more questions than it answers.2351

Energy Policy

Bulgaria’s energy planning is guided by two key documents: the Energy Strategy for 2023 to 2053 and the updated Integrated National Energy and Climate Plan (INECP) released in January 2025. The Energy Strategy, announced in January 2023 by then-acting Energy Minister Rossen Hristov, outlines long-term goals, including the elimination of coal by 2038—incidentally corresponding to Germany’s coal phaseout target date—with reductions starting in 2030, while maintaining Bulgaria’s role as a regional electricity exporter.2352 It envisions the installation of 7 GW of solar and 2 GW of wind power by 2030, increasing to 12 GW and 4 GW, respectively, by 2050, alongside investments in battery storage, electric vehicle charging stations, and additional hydropower.

Nuclear energy is a cornerstone of the Energy Strategy; it plans for 4 GW of additional nuclear capacity, including 2 GW of capacity at the Belene site—now cancelled—by 2035–2040 and an additional 2 GW at Kozloduy by 2045.2353 In comparison, the INECP specifies that total nuclear capacity is to reach 3.27 GW by 2035 and 4.53 GW in 2040 (when projections end).2354 Complementing this, the updated INECP mentions a 35 percent share of renewable energy in gross final energy consumption by 2030 and 49 percent in final electricity consumption (gross). It also emphasizes hydrogen infrastructure development, such as a 100 percent hydrogen pipeline and cross-border transport with Greece and Romania as well as energy efficiency measures like renovating buildings and constructing nearly zero-energy buildings.2355 Together, these plans aim to balance energy security, environmental sustainability, and economic growth while transitioning Bulgaria to a low-carbon economy.

In 2024, nuclear dominated by far the country’s electricity mix accounting for 41.8 percent of power generation, followed by coal at 21.6 percent. The remainder was divided between solar PV (14.4 percent), hydro (7.8 percent), natural gas (5 percent), bioenergy (4.9 percent) and wind (4 percent).2356

Bulgaria is also considering the deployment of Small Modular Reactors.

Czech Republic

The Czech Republic has six Russian-designed reactors in operation at two sites. Dukovany houses four VVER-440/V-213 reactors and Temelín operates two VVER-1000/V-320 units. The units of Dukovany were connected to the grid in the years 1985 to 1987 and both units of Temelín in the year 2000. Each plant’s operating license is unlimited but conditional on a positive period safety review every ten years. In 2024, nuclear power production stood at 28 2TWh—a slight decrease from 28.9 TWh in 2023—and represented a 40.2 percent share in total electricity generation.

In late 2024, the Czech Republic issued an update of its National Energy and Climate Plan (NECP), which is also reflected in the ČEZ Group’s annual report.2357 The ČEZ Group is a large energy utility company in the Czech Republic, which, among a range of power stations, operates the two nuclear power plants of the Czech Republic. According to the plan, by 2030, the nuclear share in electricity generation is to reach 44 percent, with a 2040 target to rise to a potential maximum of 68 percent, before decreasing again to 46 percent in 2050.2358

The NECP indicates three areas impacting nuclear capacity:

• The confirmation of the indefinite extension of the operating licenses of all existing units with an obligation to undergo an in-depth periodic safety review every ten years, see Czech Republic Focus in WNISR2024. The current estimated operating lifetime is at least 60 years.
• The construction of two to four new large units at Dukovany and possibly Temelín with the first new unit planned to start up in 2036.
• The deployment of small and intermediate modular reactors at up to three locations with a total capacity of up to 3 GW by 2050 and a first unit operating by the mid-2030s.

Nuclear Fuel—Reducing Russia Dependencies

As reported in previous issues of WNISR, the Czech Republic is looking to reduce its dependency on Russian natural uranium and fuel services including enrichment, conversion, and fuel-assembly fabrication. Further steps toward this goal were taken over the past year. Temelín and Dukovany are expanding their fresh fuel storage-capacities, which, once the modification is implemented, would be able to hold more than ten fuel loads per site.2359 A fuel delivery contract for Russian fuel ended in 2024. Alternative fuel supply from Westinghouse—initially set to begin in 2024—was delivered at both sites in May and June 2025.2360 At the beginning, fuel of mixed origins would be loaded in the cores.

In October 2024, Framatome and ČEZ signed a Memorandum of Understanding regarding “Framatome’s own-design VVER-1000 fuel development program”.2361 Earlier, in 2022, ČEZ had contracted Framatome for the supply of fuel assemblies for Temelín starting in 2024.2362 This would in principle give access to another potential non-Russian fuel supplier besides Westinghouse, however, Framatome is planning to fabricate the fuel under a Russian license agreement in Lingen, Germany. The project still needs a permit from the relevant German authorities, which means that this option is taking more time than envisaged. A first delivery is now expected in 2026.2363 The two companies have been negotiating a further agreement for fuel supply of Dukovany for a few years (see WNISR2024) and were still discussing as of May 2025.2364 (See Russia Nuclear Interdependencies.)

In November 2024, the signature of a 10-year contract with Urenco on enrichment services for Dukovany and Temelín units “until the mid 2030s” was announced, without the mention of exact dates or volumes involved.2365

In April 2025, ČEZ signed a contract with Kazakhstan’s National Atomic Company Kazatomprom JSC for the supply of natural uranium concentrates, which is planned to cover approximately one-third of Temelín’s uranium needs for Westinghouse-manufactured fuel assemblies over the following seven years.2366

New Large Reactors at the Dukovany Site

As reported in previous WNISR editions, the Czech Republic started planning for new reactors over two decades ago. Originally, new units were planned to replace the capacity lost due to decommissioning of the existing units at the Dukovany site. During a long planning and tender process, led by ČEZ Group’s Elektrárna Dukovany II (EDU II), finally three bids, provided by U.S.-Canadian Westinghouse, French Electricité de France (EDF), and Korea Hydro and Nuclear Power (KHNP), were received. See Czech Republic Focus: Newbuild Projects in WNISR2024 for a more detailed account of the process.

In June 2024, ČEZ announced KHNP—offering its APR-1000 design, derived from the System 80 design by Westinghouse and a never-built, downsized version of KHNP’s APR-1400 design—as its preferred bidder.2367 The following month, the government confirmed the selection of the Korean supplier for the construction of two units at Dukovany.2368

While the original tender was just for one unit, EDF and KHNP, had been asked to submit updated binding offers by April 2024 covering the construction of up to four reactors in the hope of reducing per unit costs.2369 KHNP’s winning offer was announced to be “around CZK200 billion [US$9.1 billion] per unit.”2370

Following the decision, Westinghouse as well as EDF filed appeals with the Czech Office for the Protection of Competition (ÚOHS), contesting the legality of several decisions along the procurement process.2371 Further, building on an earlier dispute, Westinghouse asserted that KHNP was exporting technology protected by Westinghouse property rights,2372 while EDF argued that the fixed price offered by KHNP—electricity generation reportedly for less than €90/MWh (US$102/MWh)—would not be possible without Korean public subsidies which thus would constitute unfair competition.2373 In October 2024, EDF took the additional step to file a complaint with the European Commission against the Czech decision.2374

ÚOHS put a temporary hold on the conclusion of a contract with South Korea’s KHNP then expected in March 2025.2375 In a first decision, the complaints were largely dismissed in October 2024;2376 later EDF’s claims were rejected in a second and final decision.2377

In November 2024, the South Korean and the U.S. governments signed a Memorandum of Understanding (MoU) on Principles Concerning Nuclear Exports and Cooperation.2378 A second MoU was initialed in early January 2025, finalizing “the provisional understanding reached by the participants in November 2024.”2379 Finally, a few days later, Westinghouse, KHNP, and KHNP’s parent company Korea Electric Power Corporation (KEPCO) sealed a direct final settlement agreement of the long-standing legal conflict between the two companies.2380 But while it published that an arrangement had been made, the terms of the deal were kept secret.

In May 2025, the European Commission sent a non-binding request to the Czech Republic, asking it to suspend the signing of the contract with KHNP until an investigation on whether KHNP received improper state subsidies related to the tender was concluded.2381 A leaked version of the letter laid out that there were “significant indications” that KHNP may have received “foreign subsidies that are liable to distort the internal market”, in violation of E.U. regulations and signaled that the Commission was “in the process of preparing a decision initiating an in-depth investigation.”2382 The dispute further escalated with the now majority government-owned EDU II company openly calling on EDF “to publish its offer immediately” to allow the public to judge the bids by themselves, suggesting that the French company was “not interested in winning the tender, but in ensuring that no nuclear power plant is built here in the Czech Republic at all.”2383

On 6 May 2025, the Czech Brno Regional Court prohibited the signature of the agreement with KHNP—reportedly set for the following day—until the legal case initiated by EDF against ÚOHS was fully processed in court.2384 On 4 June 2025, the Czech Supreme Administrative Court nullified the lower court’s injunction and paved the way for contracts to be finalized.2385 Prime Minister Fiala and representatives of ČEZ announced the signing of the engineering, procurement, and construction (EPC) contract for the EDU II project at a press conference the same day.2386 The agreement comprises the construction of two units of APR-1000 reactors at the Dukovany site (Units 5 and 6) for CZK 400 billion (~US$18 billion). Target dates for completion are 2036 and 2037, respectively.2387

The implementation will be challenging. Besides many technical construction issues, the prescriptive Czech regulatory system is tailored toward Russian VVER reactors, the only nuclear technology operational in the country.

Meanwhile, in April 2025, the Czech government acquired 80 percent of the ČEZ project company Elektrárna Dukovany II (EDU II).2388 The government aims for 60 percent national content of the work commissioned in the framework of the project.2389

Small and Intermediate Modular Reactors

As mentioned above, ČEZ targets SMR deployment totaling to a capacity of up to 3 GW by 2050, with the first unit to be built at the Temelín site.2390 The very optimistically envisaged target period for the start of SMR deployment is 2030–2035. The SMR strategy took a step forward in October 2024, when Rolls-Royce, whose technology had previously been selected as preferred design, announced a strategic partnership with ČEZ, materializing in ČEZ gradually acquiring an equity investment of “approximately 20 percent” in Rolls Royce’s subsidiary Rolls-Royce SMR.2391 The acquisition was sanctioned by the European Commission in February 20252392 and took effect in March 2025.2393 In addition, ČEZ contracted the company Amentum’s Brno-based branch to prepare an environmental impact assessment for two sites, one being Temelín, the other Tušimice, where a coal-fired power station was due for decommissioning.2394 The Rolls-Royce SMR design features a 470-MWe pressurized water reactor.2395

Power System

In 2024, the Czech Republic produced a total of 74 TWh of electricity (gross). With 36 TWh, fossil fueled thermal power plants were leading, followed by nuclear power plants with 30 TWh, hydro and solar with 3.6 TWh each, and wind plants with 1 TWh. Close to 7 TWh of electricity was exported. Comparing the situation of 2024 with five years ago, the generation from hydro, wind, and nuclear power was practically unchanged. There was a significant 28-percent decrease in production from fossil fuels (36 TWh in 2024, down from 50.6 TWh in 2019) and a noteworthy 55-percent increase in production from solar power (3.6 TWh in 2024, up from 2.3 TWh in 2019), which however remains at a modest level contributing ten times less than fossil fuels—or 4.9 percent— to the total. Overall electricity production declined from 87 TWh gross in 2019 to 74 TWh in 2024, which was balanced by less exports and less domestic consumption.2396

Hungary

Hungary has one operating nuclear power plant at Paks, where four VVER-440/v-213 reactors provided 16 TWh (15.2 TWh net) or 42.8 percent of the country’s electricity in 2024; whereas the production remained stable (around 0.1 TWh difference), the share in electricity generation dropped from 44.8 percent in 2023. It peaked at 53.2 percent in 2014.2397

Fuel for operating the Soviet-designed nuclear plant, so far, has been provided solely by Russian TVEL.2398 However, in late 2023, the Hungarian Parliament passed an amendment to the official nuclear policy that would allow Hungary to seek alternative nuclear fuel sources.2399 In October 2024, the Hungarian government signed a contract with Framatome for the long-term fuel supply of the four VVER-440 reactors, starting in 2027.2400 For a detailed account and earlier developments of Hungarian dependence on Russian nuclear fuel, see past WNISR editions and Russia Nuclear Interdependencies.

Paks I – Operating

Currently Paks I is licensed for 50 years of operation—following 20-year license extensions granted for each of the four reactors between 2012 and 2017—on the condition that a periodic safety review confirms required safety margins every decade for the duration of the license. In December 2023, the government officially communicated to the European Commission its intention to extend the operational lifetime of the four reactors by another 20 years, enabling operation to 70 years, i.e., until the 2050s. A detailed implementation plan shall be provided by 2028. Costs for “revamping the electric and control systems” were estimated at around €1.5 billion (US$20231.6 billion).2401

While according to previous planning, Paks I would have been decommissioned once Paks II (see below) was connected to the grid, now parallel operation of Paks I and Paks II is envisaged. This raises, amongst various issues, questions about the environmental impact to the Danube River. A limit of 30°C water temperature is imposed to protect wildlife. Due to this limit, Paks I already had to reduce power output on several days in previous years, and despite that, temperatures reportedly still exceeded the limit on several occasions.2402 In summer 2024, the government passed a decree under which, when water temperatures reach or exceed the limit, the Ministry of Energy can allow the units to continue operating at full power, if it considers reduced output or shutdown to threaten energy security,2403 and in 2025 an advisory body recommended to increase the temperature limit by 2°C.2404 While such an increase might reduce the low power days for Paks I with a combined capacity of roughly 2 GW, it is unclear how Paks I and Paks II with a combined capacity of roughly 4.4 GW would operate within those limits. Constructing cooling towers or adding air cooling that could ease the water availability constraints are not considered and would be extremely pricey.

Paks II – Planned

For a decade and a half, plans have been discussed and developed to build additional nuclear power plants. In March 2009, Parliament approved a government decision-in-principle to build additional reactors2405 and a tender was prepared according to E.U. rules. In 2014, the Paks II project, consisting of two 1200-MW VVER-1200s (Paks-5 and -6), was suddenly awarded to Rosatom without reference to any public tender, with Russia financing 80 percent of the project through loans.2406 The original 2014 Russian-Hungarian bilateral financing agreement consisted of a €10 billion (US$201413 billion) loan to the Hungarian state. Hungary would have to invest 20 percent or up to €2.5 billion (US$20143.3 billion) into the project. In April 2021, the loan terms were revised to allow Hungary to start repaying the loan in 2031, five years later than originally agreed.2407

Legal Challenges to State Aid for Paks II

In November 2016, the European Commission cleared the award of the contract to Rosatom of any infringement of its procurement rules,2408 and in March 2017, it also approved the financial package for Paks II.2409 However, in February 2018 the Austrian government filed a legal action challenging the validity of the decision.2410 In November 2022, the European General Court sided with the Commission, notably arguing that because Hungary’s state aid for the Paks II project “concerns solely the costs of investment in two new reactors intended to replace the four old reactors […] and with no operating aid being foreseen, the effect on the energy market will only be limited.”2411 The Austrian government subsequently filed an appeal. In February 2025, the European Court of Justices’ Advocate General proposed as legal solution that the court uphold the Austrian government’s appeal and set aside the earlier judgment from the General Court.2412 The court is yet to render its judgment.

The financial intergovernmental agreement and the Engineering, Procurement and Construction (EPC) contract were subject to draft amendments in 2023. According to the government, the new amendments were approved by the European Commission in May 2023,2413 allowing for the amended contracts to be signed and enforced in August 2023.2414 In fact this marked the sixth amendment of the EPC contract, as it had been revised in the past notably to extend completion dates.2415

Paks II Licensing and Construction Process

Despite the economy-wide sanctions against Russian companies, Paks II is proceeding, as Hungary obtained an exclusion of the project from potential future E.U. sanctions in June 2024, by making it the main condition for its government to approve the E.U.’s 14th sanctions package against Russia.2416

Although construction licenses for the two new VVER-1200 reactors were already granted in August 2022,2417 only site preparation works and construction of certain components was possible as the license was tied to several conditions, most notably the approval of the preliminary safety analysis report, which was not granted in 2022.

In August 2024, the core catcher for the Paks II first unit (Paks-5) was delivered to the site following a 42-day shipment by a water route organized by an Austrian company.2418 Furthermore, manufacturing of the reactor pressure vessels for Units 5 and 6 has commenced in Russia, as well as production of the Arabelle turbine for Unit 5 in Belfort, France.2419

In November 2024, the Hungarian Atomic Energy Authority (HAEA) finally approved the preliminary safety analysis report, a key condition in the construction license for starting work on the nuclear island.2420 However, the partial collapse of a berm at the pit of Unit 5 in January 2025 halted work in the area pending reinforcement measures to preserve the integrity of the structure.

In June 2025, upon reviewing the implemented requirements, the regulator authorized resumption of work in the zone, but noted that “other activities, such as the pouring of the first concrete, may only commence once the conditions and hold points set out in other HAEA permits have been met and accepted.”2421

Export of Siemens I&C for Paks II?

As reported in previous WNISR editions, export grants for the Instrumentation and Control (I&C) System, which was to be provided by Siemens Energy with Framatome, were being withheld by the German government in early 2023.2422 As retaliation, the Hungarian government threatened Siemens with the cancellation of other orders, e.g. for locomotives, and was seeking to focus on Framatome as a major European supplier for nuclear plant components.2423 See also Russia Nuclear Interdependencies on Siemens implications.

In March 2023, Péter Szijjártó, Minister of Foreign Affairs and Trade, had suggested bringing Russian suppliers into the mix if Framatome failed to “take over leadership of the [Franco-German] consortium.”2424 Reportedly, the Hungarian government had been working on sidelining Siemens to cooperate solely with Framatome.2425 However, at the opening of Siemens Energy’s new gas-turbine component manufacturing facility in Budapest in May 2024, Szijjártó said, in a surprise announcement, that the company would indeed be delivering I&C equipment to the Paks II plant through its Hungarian subsidiary.2426 Apparently, Siemens neither confirmed nor denied the information.

After the announcement of the project’s exemption from future E.U. sanctions, Paks II CEO Gergely Jákli was quoted as saying in June 2024 that he was “certain that Siemens Energy will […] fulfill its contractual obligations,” albeit noting that Germany’s Federal Office for Economic Affairs and Export Control might still block the export.2427 Meanwhile, the Minister of Foreign Affairs and Trade apparently took the exemption to mean that European companies would no longer require their domestic authorities’ permission to participate in the project.2428 Later in June 2025, it was announced that Siemens Energy was relocating its dedicated division to Budapest, in an apparent attempt to evade the need for German authorities’ consent.2429 There are, however, still doubts if this measure really could substitute application for and granting of an export license from the competent German authorities.2430

Energy Policy Developments—Solar Provides One Quarter of Power

Renewable energies have been increasing significantly over the past decade, driven by the expansion of solar PV that went from just 89 MW in 2014 to 7.7 GW in 2024, representing over 88 percent of the country’s renewable power capacity.2431 Consequentially, electricity production from solar has accelerated by a factor of 2.5 in just three years from 3.8 TWh (10.5 percent) in 2021 to 9.4 TWh (24.6 percent) in 2024.2432 According to EMBER, in 2024 Hungary had the highest share (25 percent) of solar in its electricity mix worldwide.2433

In 2024, Hungary had the highest share (25 percent) of solar in its electricity mix worldwide

Hungary’s updated National Energy and Climate Plan from October 20242434 defines a 12-GW solar PV target for 2030, while wind power capacity shall triple to a modest 1 GW. However, those numbers should be read with care. Deployment of renewables in Hungary is mainly limited by the power grid. The planned capacity expansions for wind and solar reflect planned extensions of the Hungarian grid. Grid planning took place at a time when the Hungarian legislation was unfavorable for wind power, but it has been amended since.2435 The ratio between wind and solar capacities might change; according to sources, it likely will change. Whatever the contribution of individual sources, the total renewable electricity generation share is to reach 32 percent by 2030.

The plan assumes continued operation of all four reactors at Paks I, plus the commissioning of the planned Paks II reactors, amounting to a total 4.4 GW of nuclear capacity. Accordingly, it is expected that nuclear power will account for about 20 percent of power capacity by 2040 and 18 percent of the projected 2050 mix.2436 However, Hungary is also looking into deploying Small Modular Reactors (SMRs), whose inclusion to official energy policies would affect the projected power mix.2437

According to EMBER, natural gas contributed 18.7 percent to the Hungarian electricity production in 2024. The remainder consisted of a mix of coal (6.3 percent), bioenergy (4.6 percent), hydro (0.6 percent), and “other” fossil fuels (1.2 percent).2438

Romania

Romania has one nuclear power plant at Cernavodă, where two Canadian-designed CANDU heavy-water reactors are in operation, with a total capacity of 1.3 GW. In 2024, they provided 10.4 TWh net (10.9 TWh or 20.1 percent of gross generation), slightly down from 10.6 TWh net in the previous year and just shy of the historic maximum of 20.5 percent in 2020.2439

The Romanian electricity mix is dominated by hydro that supplied 26.6 percent of the country’s power in 2024, followed by nuclear (20.4 percent2440), gas (18.6 percent), and coal (13 percent). Non-hydro renewable generation is dominated by wind that supplied 12.2 percent of the electricity in 2024. Solar PV supplied only 7.9 percent. This, however, is more than double of the 2023 share, which was only 3.9 percent.2441 The higher share in 2024 came from an addition of 1.7 GW of solar capacity. For 2025, more than 2 GW of additional solar capacity is expected.2442 This is complemented by a substantial rollout of battery storage, with the country aiming for 5 GW by the end of 2026.2443

Romania’s final National Energy and Climate Plan envisions a total of 8.2 GW of solar PV, 7.3 GW of wind, and 6.9 GW of hydro capacities by 2030, bringing the share of renewables in the electricity mix to 75 percent. The country is also planning a coal phaseout by 2032. Moreover, the plan envisions the expansion of nuclear capacities, including 1.4 GW from the slightly uprated existing reactors, 462 MW from Small Modular Reactors (SMRs) scheduled to be commissioned in 2030, and another 1.4 GW from two new 700-MW units to start up at Cernavodă in 2031 and 2032, respectively. Total operating nuclear capacity would thus rise to 3.2 GW by 2035.2444

Plans for Romania to become the first European country to deploy SMRs have been ongoing since March 2019, when U.S. company NuScale and Romanian power utility Nuclearelectrica signed a first MoU to explore the potential of SMR deployment.2445 The company RoPower—co-owned by Nuclearelectrica and Nova Power & Gas—was created specifically to develop 462 MW of SMR capacity (six 77-MW NuScale units) at the site of a former coal power plant at the Doiceşti in 2022, and Front-End Engineering and Design (FEED) studies have been ongoing since.2446 Formal approval for transitioning to FEED Phase 2 was received in late 2023,2447 and Fluor was contracted in July 2024 to carry out related activities, which are due to last 15 months2448. Refer to past editions of the WNISR for more details.

Over the course of 2024, preparatory design and licensing work continued. Building on early U.S. support to the project, U.S. EXIM (Export-Import Bank of the United States) committed a US$98 million loan in late 2024 for Phase 2 of the FEED studies; formal approval for negotiations with RoPower ensued,2449 but there is no indication as of mid-2025 that the loan agreement was finalized. While the project’s completion was initially planned for 2027 or 2028,2450 and U.S. institutions had issued letters of interest for US$4 billion in financial support early on,2451 RoPower is yet to take a final investment decision.

In a presentation of their quarterly report in May 2025, NuScale’s president and CEO stated “discussions are underway to extend the project into the detailed design phase, a critical step that would enable the submission of a final investment decision application to the Romanian Government by late Q1 or early Q2 2026,”2452 and Nuclearelectrica in its annual report still expects the first unit to be connected to the grid “by the end of this decade”, though it notes “the timeline will be fine-tuned after completion of the FEED studies…”2453

In a June 2025 announcement ahead of its subsequent general meeting, Nuclearelectrica presented its intent to bring all six modules to the grid by 2031.2454 While the planned site has been fully leased to RoPower,2455 the operation of the first SMR by 2030 seems highly unlikely given that

• the chosen 77-MW module was only recently licensed in the U.S. (May 2025)2456 and is yet to be in Romania (a “pre-license” was granted by regulatory agency CNCAN in October 2023, but a full license of the design or project permit has not been granted yet),2457
• NuScale has never built a single reactor anywhere (see chapter on SMRs), and
• the FEED study might be delayed.2458

The two operational Romanian PHWRS are the only CANDU reactors operating in Europe. Construction started between 1982 and 1987 on five units. Following years-long construction interruptions, Unit 1 was completed in 1996 and Unit 2 started up in 2007, respectively 14 and 24 years after construction originally started. Both were partly funded by the Canadian Export Development Corporation, and Unit 2 was also partly funded by the Euratom Loan Facility. Construction of the remaining three units was interrupted and, so far, has not resumed.

As with other ageing CANDU reactors, major refurbishment will be needed to ensure continued operation past the initially envisaged 30-year lifetime. In 2017, the plan to upgrade Unit 1 to allow for a 30-year lifetime extension was initiated.2459 In December 2024, an Engineering, Procurement and Construction (EPC) contract was signed between Nuclearelectrica and an international consortium consisting of the Canadian company AtkinsRéalis, Italian Ansaldo Nucleare, the Canadian Commercial Corporation, and Korea Hydro & Nuclear Power (KHNP).2460 The contract, valued at €1.9 billion (US$20242 billion),2461 was approved by the shareholders of Nuclearelectrica in January 2025,2462 and by the Canadian government in March 2025.2463

Total investment required for the refurbishment program is estimated at €3.5 billion (US$4 billion).2464 The timeline for the project envisions a temporary shutdown of the reactor from 2027 to 2029 to allow for refueling and refurbishment work to be carried out, after which it is to resume operations for 30 years. As of mid-2025, it appears the unit might now be expected to resume operation in 2030,2465 and according to the Annual Report 2024 of Nuclearelectrica, the “kick-off [remains scheduled for] Q4 2027.”2466

Additionally, various foreign companies have been involved in attempts to revive the construction of Units 3, 4, and 5 of the Cernavodă plant. For example, in November 2013, Nuclearelectrica and China General Nuclear (CGN) signed a letter of intent for the construction of Units 3 and 4.2467 Though it was followed by further preliminary arrangements, this project did not materialize. Refer to earlier editions of the WNISR for details.

Most recently, there has been some development regarding the resumption of construction work at Units 3 and 4, whose respective constructions are supposedly 52 and 30 percent complete.2468 In October 2024, the Romanian Energy Minister Sebastian Burduja said that the government was “putting this strategic project onto a path of no return, and in 2031-2032 [Romania] will have two new nuclear reactors at Cernavodă.”2469 The Limited Notice to Proceed (LNTP), the first phase of an EPC Management (EPCM) contract valued at €3.2 billion (US$20243.4 billion), was signed in November by Nuclearelectrica’s subsidiary EnergoNuclear with a joint venture consisting of Fluor, AtkinsRéalis, Ansaldo Nucleare, and Sargent & Lundy. This first part of the EPCM contract covers preliminary work for the next 24–30 months. However, the second phase of construction implementation (80–84 months), i.e. the Final Notice to Proceed (FNTP), remains “subject to commercial terms being further refined and agreed and the Final Investment Decision being taken in line with the Support Agreement between the Romanian State and SNN.”2470 Adding up the allocated time for the two phases of the contract, and assuming the fastest possible completion of the whole project, including construction, in 104 months (8.7 years), the target of operating both units by 2032 is mathematically slightly out of scope.

The project is to be financed by a complete loan guarantee by the Romanian state and has received support totaling €6.1 billion (US$7.01 billion) from the U.S. EXIM Bank, Canada, and Italy. According to Nuclearelectrica, in July 2024 the European Commission’s Directorate General for Energy provided a “positive opinion” after assessing “the technical and nuclear safety aspects of the project”.2471 There is no official estimate of the total cost of the project, but over the years, various statements drew an estimated range of “at least €6.5 billion [US$20237 billion]”2472 to US$8 billion.2473

Slovakia

Slovakia, like neighboring countries such as the Czech Republic and Hungary, is heavily reliant on nuclear energy for electricity production. Over the past 20 years, nuclear energy has continuously contributed over 50 percent of Slovakia’s electricity. This share rose to a historical maximum of 61.3 percent in 2023 (61.5 percent in 2024) following the commissioning of the Mochovce-3 reactor, and nuclear generation remained stable in 2024.2474

Slovenské Elektrárne (SE), which until May 2025 was jointly owned by the Czech energy group Energetický a Průmyslový Holding (EPH), the Italian utility ENEL (Ente Nazionale per l’Energia Elettrica), and the Slovakian government, is now co-owned by EPH (66 percent) and the Slovak state (34 percent).2475 The company operates two nuclear power plants: Jaslovské Bohunice and Mochovce. Jaslovské Bohunice houses two operational VVER-440/V-213 reactors, while the Mochovce plant, as of mid-2025, has three similar reactors connected to the grid. Together, these facilities provide a total capacity of 2.3 GW.2476 A fourth unit at Mochovce has been completed in recent years and is currently undergoing pre-commissioning testing.

The commissioning of Mochovce-3 led to a 15 percent increase in Slovakia’s nuclear electricity production as it rose from 15.9 TWh (15 TWh net) in 2022 to 18.3 TWh (17.4 TWh net) in 2023. In 2024, nuclear provided a stable 18.2 TWh (17.3 TWh net) representing 61.5 percent of the country’s total electricity generation.2477

The country has three closed reactors at the Bohunice site. The first is the A-1 reactor, a small 93-MW unit that began operation in 1972 but was closed in 1977 following a series of accidents. The other two are VVER-440/V-230 reactors, which were closed in 2006 and 2008, respectively, as part of Slovakia’s agreement to join the European Union in 20042478 (for more information see Slovakia in Decommissioning Status Report).

The operating Units 3 and 4 of the Bohunice plant (commonly known as Bohunice V2), which have been in commercial operation since 1985, underwent an extensive modernization program between 2000 and 2010. This program included capacity uprates, increasing their output from 440 MWe to 505 MWe (gross). Similarly, the capacities of Units 1 and 2 at the Mochovce plant, which began operation in 1998 and 2000, respectively, were uprated from 440 MWe to 470 MWe and eventually 500 MWe (gross).2479 SE plans to operate Bohunice Units 3 and 4 until 2044 and 2045, respectively, and Mochovce Units 1 and 2 until “at least” 2058 and 2060, respectively.2480

The Mochovce Saga

In April 2006, the Italian utility ENEL acquired a 66-percent stake in SE and committed to investing €2 billion (US$20062.5 billion) between 2006 and 2013.2481 This included funding the completion of Mochovce-3 and -4, two Soviet-era VVER-440/V-213 reactors, whose construction began in January 1985 but was halted in 1993.2482

After a long process muddled by financial issues causing disputes between the Slovak government and ENEL, Mochovce-3 finally achieved first criticality on 22 October 2022, was connected to the grid on 31 January 2023, and reached full capacity on 22 September 2023 (see section on Slovakia in WNISR2023 for a detailed account of the construction and financing history).2483 Following a 114-hour test run in October, SE announced on 17 October 2023 that the reactor’s commissioning was complete.2484

Unit 3’s grid connection, planned for 2012 when construction resumed, was delayed by another 11 years, while Unit 4, then set for 2013, faces at least a 12-year delay since construction resumed, with uncertainty around the latest 2025 target. Estimated project costs rose from €20072.8 billion (US20073.8 billion) at the 2007 relaunch to €20206.2 billion (~US$20207 billion) by December 2020.2485 In February 2025, SE closed a €3.6 billion (US$4.1 billion) Term and Revolving Facilities Agreement with a bank consortium, which SE termed “the largest corporate debt refinancing in Slovakia to date” that would “primarily be used to refinance existing senior and subordinated indebtedness and fund general corporate purposes, including working capital.” SE also stated that “completing the nuclear power plant Mochovce 4 is the Company’s primary short-term objective.”2486

In October 2024, SE confirmed the successful testing of active and passive emergency systems at Mochovce-4.2487 Cold hydrostatic testing began in December 2024,2488 followed by hot hydrostatic testing in March 20252489. These tests verify the safety, pressure resistance, and leak tightness of the primary circuit and related components. Active testing with nuclear fuel is planned next, pending approval from the Slovak nuclear regulatory authority.

SE is planning a district heating system using waste heat from the Mochovce plant to supply the town of Tlmače.2490 The status of the licensing procedure for the project is unclear.

A significant issue in Slovakia, just as on the global level, is the lack of qualified staff (welders, engineers, operators etc.) in the sector, going as far as creating the need to rehire senior workers who have already retired.2491 This issue becomes particularly acute, as observed in Slovakia, when combined with challenges in educating and training new personnel in a timely way as well as difficulties in transferring knowledge amongst an imbalanced mix of young and experienced operators. According to knowledgeable sources, the resulting shortage of experienced professionals, coupled with the only option to hire inexperienced staff, has contributed to a rise in the number of reportable events in recent years.

New Reactor for Bohunice

The Jaslovské Bohunice site has long been considered for new nuclear projects. In 2009, the Nuclear Energy Company of Slovakia (Jadrová energetická spoločnosť Slovenska or JESS), a joint venture by Jadrová a vyraďovacia spoločnosť (JAVYS, 51 percent) and ČEZ, was established for the so-called NJZ project, which envisioned the construction of a Generation III PWR with a capacity of up to 1700 MW (net).2492 In February 2023, JESS requested a siting permit for a new reactor2493 and received a permit in May 2024 for a new generation PWR with an electrical output of up to 1200 MW (gross).2494

In July 2023, JAVYS signed an MoU with Westinghouse to explore deploying both its AP-1000 and AP-300 SMR in Slovakia.2495 In May 2024, Prime Minister Fico, who took office for a fourth (non-consecutive) term in October 2023, announced plans for a 1200-MW state-owned nuclear power plant at the Bohunice site.2496

The Minister of Economy D. Sakovà stated during a press conference in November 2024 that “initial construction work should take place in April 2032, the first fuel loading in early 2038, and commissioning in 2038 as well.”2497 A statement put out by the ministry at the time mapped the same milestones and indicated that a tender was expected to be launched in 2027.2498 Reportedly, the project was then expected to cost €10 billion (US$202410.9 billion);2499 as of June 2025, a significantly inflated figure of €13–15 billion (US$14.8–17 billion) was circulating in media reports.2500 In its annual report released in March 2025, JAVYS expects commissioning by 2040.

In May 2025, Prime Minister Fico negotiated with representatives of Westinghouse in Bratislava to expand cooperation in the nuclear field including through the potential construction of an AP-1000,2501 and in June 2025, following a meeting with Economy Minister D. Saková, he stated that an intergovernmental agreement with the U.S. on the construction of a new nuclear power plant was being prepared but that “we are waiting for the final position of the USA.”2502

The situation regarding potential suppliers and the tender procedure for the planned 1200-MW nuclear power plant at Bohunice remains unclear. On the one hand, only Rosatom and EDF offer reactor designs with the announced 1200 MW electrical output, on the other hand Westinghouse officially stated that it would be interested in deploying an AP-1000 reactor in Slovakia.2503 In fact, in spite of the model name, the reference unit power of the six so far completed AP-1000s in the U.S. and China varies between 1117 MW and 1170 MW.

The ambiguity of the process extends to the tender, with conflicting reports about its timeline and format. The opposition parties SaS (Sloboda a Solidarita) and Progressive Slovakia (PS) have criticized the government’s decision to move forward with the planned 1200-MW nuclear power plant at Bohunice without a public tender. SaS expressed concerns over the lack of transparency, alleging that Westinghouse was selected as the supplier without proper scrutiny or competition. Karol Galek of SaS suggested that the process was being conducted behind closed doors, raising fears of backroom deals. PS also criticized the government for prioritizing nuclear energy over the development of renewable energy sources.2504

SMR Dreams

In September 2023, SE received a US$2 million grant from the U.S. Government’s “Project Phoenix”— a subprogram of the Foundational Infrastructure for the Responsible Use of Small Modular Reactor Technology (FIRST) program—alongside Poland and the Czech Republic, to explore converting coal-fired power plants to zero-carbon Small Modular Reactor (SMR) nuclear energy.2505 A feasibility study launched in 2024 and expected to conclude in 2025, is to evaluate five potential sites, including Nováky, Vojany, Bohunice, and Mochovce, based on technical and infrastructural criteria. Ultimately, the process excluded Nováky from consideration. As of early 2024, SE planned to complete SMR design and licensing by 2029. The timeline is ambiguous as it states for 2035 “implementation project, construction, commissioning” without clarifying what is to start and what is to be completed by then.2506

Agreements signed in summer 2023 between JAVYS, Westinghouse, and EDF also include collaboration on SMR designs like the AP-300 and Nuward.2507

In October 2024, Slovakia received a further US$5 million grant under the U.S.-led “Nuclear Expediting the Energy Transition (NEXT)” project—also under the FIRST program—which supports the implementation of SMR programs. SE CEO Branislav Strýček highlighted that SMRs are intended to complement, not replace, existing nuclear sources. The grant will fund activities such as consulting on technical and regulatory requirements, collaboration with universities and nuclear facilities, and “preparation of strategies for the introduction of SMRs.”2508

In August 2024, Slovak nuclear engineering firm VUJE partnered with U.K.-based Newcleo to advance modular reactor technologies and next-generation fuel system solutions in Slovakia. The collaboration focuses on Newcleo’s lead-cooled fast reactor (LFR) design and mixed-oxide (MOX) fuel.2509 In December 2023, an MoU between Newcleo, JAVYS, and the Slovak Ministry of Economy had already been signed, setting up the first stages of a collaboration on the deployment of so-called Advanced Modular Reactors in Slovakia.2510 By January 2025, the partnership evolved into a joint venture between Newcleo and JAVYS to develop and construct up to four LFR-AS-200 reactors at the Bohunice V1 site, with the possible assistance of VUJE during the initial feasibility study and “subsequent activities” as provided under a framework agreement signed the same day.2511 In June 2025, at a conference in Bohunice, JAVYS (51 percent) and Newcleo (49 percent) signed the Founder’s Agreement to establish the Centre for the Development of Spent Nuclear Fuel Utilisation (CVP) and VUJE and Newcleo signed a Memorandum of Understanding “regarding cooperation in the design and development of scalable data centres in connection with…AMR projects.”2512 No implementation timeline was communicated.

Russian Ties and Other Contradictions

Slovakia, like many other countries in the region, remains heavily dependent on Russian energy resources, including gas, oil, and nuclear fuel. As of April 2023, 60 percent of natural gas, 95 percent of oil, and all nuclear fuel for its VVER-440 reactors were sourced from Russia.2513 While raw oil imports declined to about 75 percent in 2024,2514 indicating gradual diversification, Slovakia continues to import Russian gas via the TurkStream pipeline and re-exports a significant portion to Austria. Efforts to reduce dependency include a deal with Azerbaijan to transport gas via Ukrainian pipelines.2515 However, long-term contracts, such as the one with Gazprom valid until 2034, and favorable Russian oil prices have reinforced this reliance, which significantly alleviates the burden on Slovakia’s budget.2516 This dependency has financial and political consequences, including strained relations with Ukraine over energy disputes and reduced profits for state-owned energy companies.

In the nuclear sector, Slovakia’s reliance on Russian technology and fuel remains significant, with record imports of 230 tons of nuclear fuel in 20232517 and 178 tons in 2024 (see Figure 55). While agreements with French and U.S. companies, such as Framatome and Westinghouse, aim to diversify nuclear fuel sources starting in 2027, the current delivery contract with Rosatom-owned TVEL was extended to 2030 (see Russia Nuclear Interdependencies).

In May 2023, SE signed an MoU with France’s Framatome to cooperate on providing “100% European” fuel for VVER reactors.2518 By July 2024, a contract was signed for Framatome to supply fuel to Bohunice and Mochovce starting in 2027. In August 2023, SE also reached an agreement with Westinghouse Electric Sweden AB, with deliveries expected one year after licensing and approval.2519 In February 2024, SE Chairman Branislav Strýček stated at Westinghouse’s 5th Annual VVER Forum that he “[expected] to finish the licensing process by 2026-2027.” 2520

However, the government noted in early 2025 that:

Only some spare parts from alternative suppliers are available on the market, such as several firms in Slovakia, the Czech Republic. Most spare parts are from the Russian Federation. SE, a.s. currently only uses spare parts and maintenance services from suppliers from the Russian Federation, which are critical for the safe operation of their nuclear facilities.2521

The slow pace of diversification is compounded by the pro-Russian stance of Prime Minister Robert Fico’s administration, which has prioritized economic gains over energy supply independence. This approach undermines E.U. sanctions on Russia and complicates the E.U.’s goal of reducing reliance on Russian energy, including ending all Russian gas and oil imports by the end of 2027.2522 Additionally, the Fico government has shifted Slovakia’s foreign policy, halting military aid to Ukraine and aligning with Hungary in blocking E.U. initiatives, further highlighting the political implications of its energy supply dependency.2523 The construction of a new VVER would cement that dependency for decades.

Energy Policy Outlook

As previously indicated, Slovakia heavily depends on nuclear energy for electricity generation with the second largest nuclear share in the electricity production mix in the world in 2024 (behind France). Hydropower accounted for around 16.4 percent, natural gas 9.2 percent, bioenergy 5.1 percent, and coal 1.8 percent of total electricity production. Solar photovoltaics made up just 2.3 percent of the electricity mix.2524 Wind energy so far hardly exists in the mix.

In April 2025, the final version of the Integrated National Energy and Climate Plan 2021–2030 was submitted to the European Commission.2525 Based on the report, over the past decade, Slovak solar PV capacity has steadily increased, reaching approximately 850 MW by 2023. By 2024, the installed capacity exceeded 1 GW, with an annual installation rate of 100 MW expected to continue. Projections indicate that by the end of 2025, the installed capacity will reach 1.2 GW, with an estimated electricity production of approximately 1.3 TWh.2526 Expansion is also planned for the wind sector with a capacity target of 750 MW by 2030, producing an estimated 1.5 TWh of electricity annually.2527

Ján Karaba, the executive director of SAPI (Slovak Sustainable Energy Association), highlighted the challenges facing renewable energy in Slovakia, stating that “despite promises from the Ministry of Economy and the Ministry of Investment and Development to increase funding for this program from EU funds, it is still unclear when this will happen. Experience shows that it could take months, or even a year.”2528

Meanwhile, the head of the Slovak energy regulator, Jozef Holjenčík, has been a long-time critic of renewables, describing them as “unreliable” and their development as “chaotic”, while also warning against “speculators” and “political promoters of green ideology”. Such attitudes and delays make it more difficult for Slovakia to achieve the E.U. climate target of a national 23 percent share of renewable energy by 2030, a goal that is already significantly lower than the E.U.’s overall target of 42.5 percent.2529

Slovenia

Slovenia’s electricity production is roughly divided among three main sources as of 2024: nuclear energy at 34 percent (8 percent below the historic maximum of 42 percent in 2022), hydropower at 31 percent, and coal at 20.3 percent. Fossil fuels made up 23.8 percent of the national power mix. Solar came in at 8.9 percent.2530

JEK

Slovenia operates a single nuclear power plant, Krško, which it jointly owns with Croatia. The 688-MW Westinghouse Pressurized Water Reactor (PWR) is referred to as JEK (Jedrska elektrarna Krško) in Slovenia and NEK (Nuklearna elektrarna Krško) in Croatia. In 2024, the plant produced 5.8 TWh of electricity (gross) or 34 percent of total generation. This marked a slight decrease from 35.3 percent in 2023, despite a marginal increase of about 0.2 TWh (gross) in absolute nuclear generation compared to the previous year.2531

The Krško reactor, operational since 1983 with an initial 40-year license, underwent a refurbishment program in 2012,2532 and its license was extended to 60 years in 2015. Despite Austrian opposition2533 (see previous WNISR editions), Slovenian authorities approved an additional 20 years of operation in 2023, allowing the plant to run until 2043.2534 The IAEA had previously carried out a pre-SALTO mission (Safety Aspects of Long Term Operation) in 20212535 and conducted a SALTO mission in 2025.2536 After the latest mission, the team recommended that the operator further develop a systematic approach for overseeing the long-term operation program; adequately document the methodology and results for identifying relevant systems, structures, and components (SSCs) for ageing management; and complete and fully document the ageing management of electrical and instrumentation and control systems, structures, and components (I&C SSCs).2537

JEK2

In 2006, the Slovenian government proposed adding a new plant at the Krško site,2538 with GEN Energija beginning feasibility studies for two 1100-MW reactors,2539 initially estimated at €2 billion (US$20062.5 billion).2540

By 2010, plans shifted to a single 1100–1600-MW reactor with a target grid connection by 2025.2541 However, by 2015, limited progress had delayed the scheduled connection date to 2030.2542

Slovenia’s Long-Term Climate Strategy Until 2050, adopted in 2021, set a 43 percent renewables target by 2030—increased to 55 percent in the 2025 NECP update—and called for a decision on new nuclear development by 2027, including Small Modular Reactors (SMRs) as an option.2543 This led the Ministry of Infrastructure to issue an “energy permit” for JEK2 shortly after in 2021.2544

In 2022, Prime Minister Golob pledged a referendum on the newbuild project once a technology is selected, aiming for broad national consensus.2545 By July 2023, GEN Energija still planned a final decision by late 2027 having updated its JEK2 application in June 2023 to accommodate a larger capacity of up to 2400 MW, hoping to “expand the range of potential suppliers.”2546

In May 2024 a parliamentary resolution on the “long-term peaceful use of nuclear energy in Slovenia” was adopted. The resolution emphasizes the supposed necessity of the JEK2 project and the deployment of SMRs for the decarbonization of the Slovenian power system.2547

The same month, alongside adopting a resolution on the “peaceful use of nuclear energy,” the Slovenian Parliament voted to hold the referendum with rare bipartisan support (71 votes for, 6 against). Critics, including the Levica (“Left”) party, argued the referendum amounted to a “blank cheque” for a potentially unviable project.2548

However, the majority of parliamentary parties, including the governing Freedom Movement and the opposition party, Slovenian Democratic Party (which initially tabled the proposal), decided to revoke the consultative referendum initiative.2549 They attributed this decision to a loss of public trust due to controversies, legal challenges, and criticism raised by environmental groups and experts over the legality of the process and the formulation of the question put to referendum, and they identified the need for more comprehensive public information, as emphasized by President Nataša Pirc Musar.2550

Though planning for the referendum was cancelled, it is expected to be implemented in 2028 at the latest, before the final investment decision, according to Prime Minister Golob.2551 The government remains committed to the project and the majority party is working on framework legislation pertaining to it. Following the referendum’s cancellation, GEN Energija announced it was continuing work on the JEK2 project while exploring the potential use of SMR technology in Slovenia.2552

In May 2024, GEN Energija presented a study suggesting that a 1300-MW reactor would be the optimal size. Estimated overnight construction costs ranged from €9.3 billion (US$202410.1 billion) for a 1000-MW reactor to €15.4 billion (US$202416.7 billion) for a 1650-MW reactor. GEN Energija is aiming for a final investment decision for JEK2 by 2028, construction to begin in 2032, and grid connection by 2039.2553

Meanwhile the JEK2 project is moving forward with two remaining potential technology vendors, EDF and Westinghouse, following KHNP’s withdrawal from the bidding competition in January 2025.2554 Both vendors have signed contracts to prepare feasibility studies, with each provider conducting its own study. The total estimated value of the two studies is €8.3 million (US$9.4 million), and they are expected to be completed by the third quarter of 2025.2555

On 1 July 2025, the announcement of the “initiative for the drafting of a National Spatial Plan (NSP) for the construction of a new nuclear power plant in Krško” was published by the Ministry of Natural Resources and Spatial Planning.2556 A public consultation was scheduled to last from 1 July to 30 September 2025.2557

Slovenia’s Dependency on Coal

Slovenia is targeting to phase out coal entirely by 2033, with significant reductions in coal use by 2030, particularly through scaling down operations at the Šoštanj thermal power plant.2558 This transition scenario is based on the expansion of renewable energy sources, optimistic growth in nuclear energy output, and the establishment of 500 MW of gas-based capacity that can later transition to renewable fuels like hydrogen. While a 100-percent renewable energy scenario has been modeled for 2050, its characteristics remain somewhat obscure in the National Energy and Climate Plan (NECP) outline.2559

Slovenia’s updated NECP, submitted in 2025, aims at climate neutrality by 2050, with a 55 percent reduction in GHG emissions by 2033, aligning with the E.U.’s Fit for 55 package. It sets specific renewable energy goals, including a 33 percent share in end-use energy, 55 percent in electricity production, 45 percent in heating and cooling, and 26 percent in transport by 2030.2560

Former Soviet Union

Armenia

Armenia has practically no fossil fuel resources, especially when compared to its neighbors Azerbaijan and Iran.2561 Yet, in 2022, natural gas—all imported—represented over 60 percent of its primary energy supply.2562 Besides natural gas, Armenia relies heavily on nuclear energy, and together the two accounted for about 70 percent of its electricity production in 2024.2563 Most of the natural gas is imported from Russia2564 via the North Caucasus–Transcaucasia Pipeline. Armenia’s efforts to trade electricity for Iranian gas have been limited by its inability to compete with Russian gas prices.

Armenia still operates a single reactor, in operation since 1980, at the Metsamor nuclear power plant (also known as the Armenian Nuclear Power Plant, or ANPP), located approximately 30 km from the capital, Yerevan. In 2024, the reactor generated 2.6 TWh of electricity, accounting for 30.8 percent of the country’s total electricity production and meeting slightly less than one-third of its electricity demand. This output has remained relatively stable since 2022. In 2024, about 30 percent of Armenia’s electricity was produced from renewable sources, primarily hydropower that produced around 20 percent, although solar energy has steadily grown to almost 10 percent. The remaining 40 percent was generated from natural gas.2565

Nuclear energy in Armenia faces significant geological and geopolitical challenges. Located in a seismically active region, the country is vulnerable to earthquakes.2566 The collapse of the Soviet Union triggered the Nagorno-Karabakh conflict with Azerbaijan, leading to two wars. Geopolitically, Armenia has been historically dependent on Russia but is now “pursuing deeper ties with the European Union, strengthening cooperation with the United States, and seeking normalization with Turkey.”2567 In March 2025, a bill to launch the process of accession to the E.U. was officially signed into law.2568

The Metsamor reactor, a first-generation Soviet-designed VVER-440/v-270, is considered to have the least adequate safety systems among operating VVER designs. Similar reactors from that era in other countries have since been closed. Following a major earthquake in December 1988, Armenia closed its two reactors in March 1989. However, severe power shortages during the early 1990s, caused by a post-Soviet energy blockade amid the First Nagorno-Karabakh War, led to the decision to reopen Unit 2 in 1993. The unit resumed operation in 1995.2569

The closure of Metsamor has been repeatedly delayed due to stalled plans for a replacement (see past WNISR editions). In 2011, the Armenian Nuclear Regulatory Authority extended the reactor’s license to 2021, and required annual safety checks starting in 2016.2570 In 2012, the government announced plans to operate it until 2026.2571 Upgrades completed in 2021 extended its lifespan and increased its capacity from 407.5 MW to 448 MW (gross).2572

After the lifetime extension project was approved and related contracts were sealed in late 2023, Rosatom began work in 2024 to further extend the unit’s lifetime to 2036 by testing the reactor pressure vessel material, assessing the effects of ageing on its strength and brittleness. Annealing of the pressure vessel was undertaken during the process to extend the reactor’s life until 2026, but further tests are to be carried out.2573

The construction of a new reactor is also fraught with geopolitical tensions. Russia is keen to maintain and expand its nuclear relationship with Armenia.2574 At the same time, the U.S. is pushing for a shift away from Russian influence. In January 2025, the U.S., still led by the Biden administration, and the Armenian government extended their ties by signing a Charter on Strategic Partnership.2575 This partnership does not include military security guarantees but focuses on energy security. Negotiations were initiated towards a so-called “123 Agreement”, which enables a cooperation in the field of civil nuclear energy.2576 In April 2025 the U.S. Ambassador to Armenia discussed the possibilities of Armenia’s nuclear sector based on the new strategic partnership with representatives from Westinghouse.2577 In response to Armenia’s growing ties with the U.S., the Russian Ambassador to Armenia, in a May 2025 visit to ANPP, stated, “Rosatom is ready to offer Armenian colleagues nuclear power plants of Russian design with reactors of various capacities, distinguished by enhanced safety and economic efficiency.”2578

In June 2023, the Armenian government established a working group to explore various options for nuclear newbuild, including Small Modular Reactors (SMRs).2579 In January 2024, Prime Minister Nikol Pashinyan, described the option of modular reactors as “politically interesting” and stated that “the accident of the latter does not require emergency response zones. In other words, if there is an accident, we will not apply measures to protect the population within a radius of a hundred kilometer.”2580 However, the credibility of such prospects has drawn some skepticism, such as from Armenia’s former Deputy Minister Hakob Vardanyan who reportedly stated in February 2025 that “the likelihood of constructing a modular nuclear power plant in Armenia is extremely low, as such plants have not yet been built anywhere in the world.”2581

As of mid-2025, the vendor and technology for the project remain undecided,2582 and the Armenian government has reportedly declined to provide any updates on the matter.2583

However, in March 2025, the establishment of a new state-owned company for the construction of the nuclear power plant was accelerated. Armenian Minister of Territorial Administration and Infrastructure, David Khudatyan stated:

The government has established a closed joint-stock company to develop a proposal for a new nuclear power plant model. It will later oversee the technical requirements and supervision before handing it over to operators. The staffing structure will soon be approved, and we are simultaneously working on extending the lifespan of the existing plant.2584

Nuclear energy is not the only option under consideration, as there is a growing emphasis on renewable energy. Solar energy is on the rise, and solar power generation has increased sixfold since 2020.2585 Armenia’s updated Nationally Determined Contribution (NDC), submitted to the United Nations Framework Convention on Climate Change (UNFCCC) Secretariat in 2021, states a 2030-target of at least 15 percent renewable energy, with plans to deploy 1 GW of solar capacity.2586 The government’s 2021 energy sector development plan underscores the importance of renewables, stating:

The fact that the solar and wind technologies are considered as part of the least cost solution for new generation under all scenarios stresses the importance to Armenia of ensuring a policy and institutional environment that supports development of these technologies to the maximum extent possible, not only to ensure the lowest cost generation but also to minimize reliance on other imported energy sources and to strengthen Armenia’s energy security and competitiveness.2587

In June 2024, the World Bank approved a US$40 million loan to help advance the country’s energy transition by upgrading its power grid.2588 However, it remains uncertain whether the recent nationalization of the electric network marks the first step toward modernizing the system. Economists have raised concerns that nationalizing the power grid could harm the investment climate. As Haykaz Fanyan, head of the Armenian Centre for Socio-Economic Studies, a Yerevan-based think tank, explained: “Initiating a process of forced re-nationalization, especially without compensation, will inevitably damage Armenia’s investment attractiveness.”2589

Belarus

Belarus, along with the United Arab Emirates, is one of only two newcomer countries to connect its first nuclear power plant to the grid in the past 14 years since Iran started up its first commercial reactor in 2011.

The contract to build the Belarusian nuclear power plant—also known as Ostrovets, after the name of the nearby town—with two VVER-1200 units, was signed with Russia’s Atomstroyexport in 2011.2590 Two years later, in November 2013, first concrete was poured for Unit 1. Construction of Unit 2 began six months later, in April 2014. The original plan envisioned commissioning the units in 2019 and 2021, but they were delayed by two and almost three years, respectively. The two units were connected to the grid in November 2020 and May 2023 and entered commercial operation in June 2021 and November 2023.2591

Before the launch of the first unit of the Belarusian plant, more than 97 percent of Belarus’s electricity was generated by gas-fired power plants over the decade between 2010 and 2019. In 2019, total electricity generation stood at around 40.5 TWh. The commissioning of the first reactor primed a continuous growth of nuclear in the electricity mix. By the end of 2024, the nuclear plant was generating 36.5 percent of the country’s electricity, while total national generation increased to 45.3 TWh. Electricity generation from gas-fired power plants declined in absolute terms by 12.2 TWh—from 39.2 TWh in 2019 to 27 TWh in 2024. Renewable energy sources (including hydro) account for about 2 percent of the national electricity mix.2592

Even before the second unit of the nuclear power plant was completed, senior Belarusian officials began discussing the possibility of building a third unit at the existing site2593 or constructing a second plant at a different location. Belarusian leader Alexander Lukashenko reconfirmed the existence of such plans in March 2025,2594 when he reportedly said, “this [the expected growth in power demand] is why we’ve asked Russians to build the second nuclear power plant if possible. We will build it on our own. Except for the reactor…We have specialists.” However, such plans have been under discussion since the very beginning of Ostrovets’ construction—as early as 20122595—but no concrete details have emerged so far.

In 2011, Russia agreed to lend Belarus up to US$10 billion for 25 years to cover 90 percent of its first plant’s expected costs. The remaining 10 percent was to be financed by Belarus. According to the agreement, the funds from the Russian loan were to be used exclusively to pay for goods, works, and services of Russian origin used in the construction of the nuclear power plant.2596 In 2021, citing delays in the construction of the plant, Lukashenko negotiated easier loan terms; the start of debt repayments was postponed by two years from 2021 to 2023,2597 and the loan interest rate was reportedly reduced from 4.11 percent to 3.3 percent.2598 In 2023, Belarus secured another loan revision, extending disbursements through 2023 and delaying repayments to April 2024. Initially, repayments under the loan could be made in either US dollars or Russian rubles. However, following the signing of the supplementary agreement in 2023, all settlements were switched to Russian rubles only.2599

It is difficult to assess the total investment, especially given the fluctuations in exchange rates and commercial confidentiality regarding actual payments. In 2019, Lukashenko said that the actual price of the nuclear plant would be below US$7 billion. However, he declined to name the exact figure, citing commercial confidentiality.2600 Speaking about the possibility of building a second nuclear power plant in the country, Lukashenko also repeatedly said that it could be partially financed using saved credit funds.2601

In 2023, the Russian Ministry of Finance stated in an explanatory note to the new loan agreement that the total amount of credit used would be around US$5.36 billion2602 and total budget revenues from the loan would be US$6.65 billion, likely accounting for interest.2603 In August 2024, Lukashenko was also quoted as saying that only US$5.36 billion from the loan had actually been used.2604 In 2020, the head of Rosatom stated in an interview that the company was working on the Belarusian nuclear project with minimal profit, and that despite the contract being denominated in US dollars, settlements were made in Russian rubles.2605

Russia and Belarus have signed several agreements on managing radioactive waste and spent nuclear fuel. In 2019, Belarus adopted a national strategy that identified reprocessing in Russia with the return of waste as the preferred spent fuel management option. The total cost of this approach was estimated at US$2.5–3.5 billion over the plant’s lifetime (up to 100 years).2606 A 2022 bilateral agreement formalized this scenario,2607 and in 2023, Rosatom subsidiary TVEL agreed2608 to help Belarus develop infrastructure for final waste disposal.

The Belarusian plant became Rosatom’s first export project featuring VVER-1200 reactors, following the construction of four such units in Russia. The reference design was Leningrad II, with two VVER-1200/V-491 units, connected to the grid between 2018 and 2020.2609

Nevertheless, several incidents and irregular situations occurred during the construction and commissioning phases in Belarus. In 2016, the reactor pressure vessel at Unit 1 was dropped during installation and had to be replaced. Eventually, a vessel, originally meant for Russia’s never completed Baltic plant, was shipped to Belarus.2610 Following the rushed commissioning, Unit 1 experienced numerous failures of transformers2611 and the turbogenerator. The commissioning of Unit 2 was also delayed by more than a year due to contamination of the primary circuit with ion-exchange resins in 2022.2612

During the plant’s construction, Lithuania raised concerns under the Espoo Convention, whose parties in 2019 concluded that Belarus provided adequate technical data but failed to justify the site choice.2613 Lithuania has consistently demanded improvements to the plant’s quality and safety in recent years, sending letters to both Belarus and the IAEA, as well as the E.U., calling for the suspension of the plant’s operation until all issues are resolved.2614

Moreover, the Belarusian regime does not foster trust through access to information in complex and high-risk matters such as nuclear projects. The E.U. has “imposed a series of sanctions against Belarus in response to the country’s fraudulent elections of 2020, its human rights abuses, and its complicity with Russia’s military aggression against Ukraine.” These sanctions are directed against individuals and specific sectors, primarily related to financial transactions and the import-export of certain products, including coal and crude oil.2615

However, moving towards isolating Belarus is not a strategy universally adopted in the E.U. In April 2023, Hungary and Belarus signed a Memorandum of Cooperation in the Field of Nuclear Energy.2616 This was followed in 2024 with the signature of a Roadmap for Cooperation between the Belarusian Nuclear Power Plant and its Hungarian counterparts. Another roadmap for 2025–2027 was signed in May 2025, as Hungary strived for Belarus’ support on the completion of Rosatom’s Paks II newbuild project.2617 The strengthening of this cooperation was also supported by Rosatom’s appointment of Vitaly Polyanin as director of the Paks II construction project in November 2023; prior to this, he led the construction of the Belarusian plant.2618

In response to the war in Ukraine, electricity imports from Russia into E.U. member states have come under scrutiny, and in May 2022, the power exchange Nord Pool decided to stop trading Russian electricity.2619 Further Baltic grid synchronization with the rest of Europe will physically exclude the import of electricity from Belarus. In February 2025, Estonia, Latvia, and Lithuania successfully connected to continental European electricity grids.2620 The loss of access to the Western European power market is significant, as higher revenues from the sale of electricity to the West improved the profitability of Ostrovets.

Russia and Belarus seek to develop a unified power market, and in December 2024, Putin and Lukashenko signed a treaty laying out the legal framework for the operation and regulation of a common market.2621 According to the Belarusian Minister of Energy, the unified electricity market will be developed in stages. The rules for the market’s operation were expected to be finalized and approved in the first half of 2025, with all interactions to be managed through authorized grid operators from both countries. However, as of July 2025, there have been no updates on progress toward this milestone.

Russia

See Focus Countries – Russia Focus.

Ukraine

See Focus Countries – Ukraine Focus.